WO2020184964A1

WO2020184964A1 - Method and apparatus for video signal processing for inter prediction

Info

Publication number: WO2020184964A1
Application number: PCT/KR2020/003349
Authority: WO
Inventors: 장형문
Original assignee: 엘지전자 주식회사
Priority date: 2019-03-11
Filing date: 2020-03-11
Publication date: 2020-09-17

Abstract

Embodiments of the present specification provide a method for video signal encoding and decoding for inter prediction. The decoding method according to an embodiment of the present specification comprises the steps of: acquiring a symmetric motion vector difference (SMVD) flag of a current block, wherein the SMVD flag is related to whether a second motion vector difference (MVD) for a second direction prediction is determined on the basis of a first MVD for a first direction prediction; identifying a condition for a bi-directional optical flow (BDOF) on the basis of the SMVD flag; and generating a prediction sample of the current block on the basis of the condition for the BDOF. In the method, an SMVD is considered as a condition for a BDOF so that an increase in coding complexity and processing time of an SMVD-coded block can be prevented.

Description

Video signal processing method and apparatus for inter prediction

Embodiments of the present specification relate to a video/video compression coding system, and more particularly, to a method and apparatus for performing inter prediction in an encoding/decoding process of a video signal.

Compression coding refers to a series of signal processing techniques for transmitting digitized information through a communication line or storing it in a format suitable for a storage medium. Media such as video, image, and audio may be subject to compression encoding. In particular, a technique for performing compression encoding on an image is referred to as video image compression.

Next-generation video content will be characterized by high spatial resolution, high frame rate, and high dimensionality of scene representation. In order to process such content, it will bring a tremendous increase in terms of memory storage, memory access rate, and processing power.

Inter prediction is a method of performing prediction on a current picture by referring to reconstructed samples of another picture. In order to improve the accuracy of inter prediction, techniques for refine a motion vector or a prediction sample are being discussed.

An embodiment of the present specification provides a method and apparatus for performing prediction based on a condition for applying a bi-directional optical flow (BDOF) and a BDOF condition when bi-directional prediction is applied in an encoding/decoding process of information for inter prediction. to provide.

The technical problems to be achieved in the embodiments of the present specification are not limited to the technical problems mentioned above, and other technical problems that are not mentioned are those of ordinary skill in the technical field to which the embodiments of the present specification belong from the following description. Will be clearly understood.

Embodiments of the present specification provide a method of encoding and decoding a video signal for inter prediction. A decoding method according to an embodiment of the present specification includes obtaining an SMVD flag of a current block, wherein the symmetric motion vector difference (SMVD) flag is based on a first motion vector difference (MVD) for first direction prediction. It relates to whether or not a second MVD for second direction prediction is determined, confirming a condition for a bi-directional optical flow (BDOF) based on the SMVD flag, and the current based on the condition for the BDOF. Generating a predicted sample of the block.

In an embodiment, when the SMVD flag is 0 and other conditions are satisfied, the condition for the BDOF is determined to be satisfied, and when the SMVD flag is 1, it may be determined that the condition for the BDOF is not satisfied.

In an embodiment, when the condition for the BDOF is satisfied, a prediction sample of the current block is a first prediction sample according to the first direction prediction in the current block, a second prediction sample according to the second direction prediction, and It is determined based on the BDOF offset, and the BDOF offset is based on a first horizontal gradient and a first vertical gradient for the first prediction sample, and a second horizontal gradient and a second vertical gradient for the second prediction sample. Can be derived.

In an embodiment, if the condition for the BDOF is not satisfied, the prediction sample is a weighted sum of a first prediction sample according to the first direction prediction and a second prediction sample according to the second direction prediction in the current block. Decoding method, characterized in that determined based on.

In an embodiment, when the SMVD flag is 1, the size of the second MVD may be the same as the size of the first MVD, and the sign of the second MVD may be opposite to the sign of the first MVD.

In an embodiment, if the SMVD flag is 1, the first reference picture for the first direction prediction is a previous reference picture closest to the current picture in display order in the first reference picture list for the first direction prediction. Is determined, and the second reference picture for second direction prediction may be determined as a next picture closest to the current picture in the display order in the second reference picture list for the second direction prediction.

An encoding method according to an embodiment of the present specification includes generating prediction information for a current block, and generating a prediction sample of the current block based on the prediction information. The prediction information includes an SMVD flag related to whether a second MVD for second direction prediction is determined based on a first motion vector difference (MVD) for first direction prediction, and generates a prediction sample of the current block The step of performing includes checking a condition for a bi-directional optical flow (BDOF) based on the SMVD flag, and generating a prediction sample of the current block based on the condition for the BDOF.

A decoding apparatus according to an embodiment of the present specification includes a memory storing the video signal, and a processor coupled to the memory and processing the video signal. The processor obtains an SMVD flag of the current block, wherein the SMVD flag is related to whether a second MVD for second direction prediction is determined based on a first motion vector difference (MVD) for first direction prediction, and , Based on the SMVD flag, a condition for a bi-directional optical flow (BDOF) is checked, and a prediction sample of the current block is generated based on the condition for the BDOF.

An encoding apparatus according to an exemplary embodiment of the present specification includes a memory for storing the video signal, and a processor coupled to the memory and processing the video signal. The processor is configured to generate prediction information for the current block, and to generate a prediction sample of the current block based on the prediction information. The prediction information includes a symmetric motion vector difference (SMVD) flag related to whether a second MVD for second direction prediction is determined based on a first motion vector difference (MVD) for first direction prediction, and the prediction In order to generate a sample, the processor is configured to check a condition for a bi-directional optical flow (BDOF) based on the SMVD flag and to generate a predicted sample of the current block based on the condition for the BDOF. .

Further, an embodiment of the present specification provides a non-transitory computer-readable medium storing one or more instructions. The one or more instructions cause a video signal processing apparatus to obtain an SMVD flag of a current block, wherein the SMVD flag is a second direction prediction based on a first motion vector difference (MVD) for first direction prediction. It is related to whether a second MVD for is determined, checks a condition for a bi-directional optical flow (BDOF) based on the SMVD flag, and generates a prediction sample of the current block based on the condition for the BDOF To control the video signal processing device.

In addition, one or more instructions according to an embodiment of the present specification generate prediction information for a current block and control to generate a prediction sample of the current block based on the prediction information. The prediction information includes a symmetric motion vector difference (SMVD) flag related to whether a second MVD for second direction prediction is determined based on a first motion vector difference (MVD) for first direction prediction, and the prediction In order to generate a sample, the one or more instructions check a condition for a bi-directional optical flow (BDOF) based on the SMVD flag, and determine a predicted sample of the current block based on the condition for the BDOF. Control the video signal processing device to generate.

According to an embodiment of the present specification, it is possible to prevent an increase in coding complexity and processing time by determining whether a condition for BDOF is satisfied in consideration of symmetric motion vector difference (SMVD) and performing prediction.

The effects obtained in the embodiments of the present specification are not limited to the above-mentioned effects, and other effects not mentioned are clearly to those of ordinary skill in the art to which the embodiments of the present specification belong from the following description. It will be understandable.

The accompanying drawings, which are included as part of the detailed description to aid in understanding of the present specification, provide embodiments of the present specification, and describe technical features of the present specification together with the detailed description.

1 shows an example of an image coding system according to an embodiment of the present specification.

2 shows an example of a schematic block diagram of an encoding apparatus in which encoding of a video signal is performed according to an embodiment of the present specification.

3 shows an example of a schematic block diagram of a decoding apparatus for decoding an image signal according to an embodiment of the present specification.

4 shows an example of a content streaming system according to an embodiment of the present specification.

5 shows an example of a video signal processing apparatus according to an embodiment of the present specification.

6 illustrates an example of a picture division structure according to an embodiment of the present specification.

7A to 7D illustrate an example of a block division structure according to an embodiment of the present specification.

8 shows an example of a case in which the ternary tree (TT) and the binary tree (BT) are divided according to an embodiment of the present specification.

9 is an example of a flowchart for encoding a picture constituting a video signal according to an embodiment of the present specification.

10 is an example of a flowchart for decoding a picture constituting a video signal according to an embodiment of the present specification.

11 illustrates an example of a hierarchical structure for a coded image according to an embodiment of the present specification.

12 is an example of a flowchart for inter prediction in an encoding process of a video signal according to an embodiment of the present specification.

13 illustrates an example of an inter prediction unit in an encoding device according to an embodiment of the present specification.

14 is an example of a flowchart for inter prediction in a process of decoding a video signal according to an embodiment of the present specification.

15 illustrates an example of an inter prediction unit in a decoding apparatus according to an embodiment of the present specification.

16 illustrates examples of spatial neighboring blocks used as spatial merge candidates according to an embodiment of the present specification.

17 is an example of a flowchart for configuring a merge candidate list according to an embodiment of the present specification.

18 is an example of a flowchart for configuring a motion vector predictor (MVP) candidate list according to an embodiment of the present specification.

19 illustrates an example in which a symmetric motion vector difference (MVD) mode according to an embodiment of the present specification is applied.

20 illustrates an example of affine motion models according to an embodiment of the present specification.

21A and 21B illustrate examples of motion vectors for each control point according to an embodiment of the present specification.

22 shows an example of a motion vector for each subblock according to an embodiment of the present specification.

23 is an example of a flowchart for configuring an affine merge candidate list according to an embodiment of the present specification.

24 shows examples of blocks for deriving an inherited affine motion predictor according to an embodiment of the present specification.

25 illustrates an example of control point motion vectors for deriving an inherited affine motion predictor according to an embodiment of the present specification.

26 shows an example of blocks for deriving a constructed affine merge candidate according to an embodiment of the present specification.

27 is an example of a flowchart for configuring an affine MVP candidate list according to an embodiment of the present specification.

28A and 28B illustrate examples of spatial neighboring blocks used in adaptive temporal motion vector prediction (ATMVP) and a sub-coding block (CU) motion field derived from spatial neighboring blocks according to an embodiment of the present specification. Shows.

29 illustrates an example of a merge mode with MVD (MMVD) search point according to an embodiment of the present specification.

30 shows an example of a decoder side motion vector refinement (DMVR) method based on bilateral template matching according to an embodiment of the present specification.

31 shows an example of a DMVR process according to an embodiment of the present specification.

32 illustrates an example of a viral matching method according to an embodiment of the present specification.

33 illustrates an example of an integer pixel search and a half pixel search according to an embodiment of the present specification.

34 illustrates an example of an extended area of a CU for bi-directional optical flow (BDOF) according to an embodiment of the present specification.

35 illustrates an example of a video signal encoding procedure for inter prediction according to an embodiment of the present specification.

36 illustrates an example of a video signal decoding procedure for inter prediction according to an embodiment of the present specification.

Hereinafter, preferred embodiments according to the present specification will be described in detail with reference to the accompanying drawings. The detailed description to be disclosed below with the accompanying drawings is intended to describe exemplary embodiments of the present specification, and is not intended to represent the only embodiments in which the present specification may be practiced. The detailed description below includes specific details to provide a thorough understanding of the present specification. However, one of ordinary skill in the art appreciates that the present specification may be practiced without these specific details.

In some cases, in order to avoid obscuring the concept of the present specification, well-known structures and devices may be omitted or illustrated in a block diagram format centering on core functions of each structure and device.

In addition, the terms used in the present specification have been selected from general terms that are currently widely used as far as possible, but in specific cases, terms arbitrarily selected by the applicant will be used for description. In such a case, since the meaning of the term is clearly described in the detailed description of the relevant part, it should not be interpreted simply by the name of the term used in the description of the present specification. .

Specific terms used in the following description are provided to aid understanding of the present specification, and the use of these specific terms may be changed in other forms without departing from the technical spirit of the present specification. For example, signals, data, samples, pictures, frames, blocks, etc. may be appropriately replaced and interpreted in each coding process.

Hereinafter, in this specification, a'processing unit' means a unit in which an encoding/decoding process such as prediction, transformation, and/or quantization is performed. Also, the processing unit may be interpreted as including a unit for a luma component and a unit for a chroma component. For example, the processing unit may correspond to a block, a coding unit (CU), a prediction unit (PU), or a transform unit (TU).

Further, the processing unit may be interpreted as a unit for a luminance component or a unit for a color difference component. For example, the processing unit may correspond to a coding tree block (CTB), a coding block (CB), a PU, or a transform block (TB) for a luminance component. Alternatively, the processing unit may correspond to CTB, CB, PU or TB for the color difference component. Further, the present invention is not limited thereto, and the processing unit may be interpreted as including a unit for a luminance component and a unit for a color difference component.

Further, the processing unit is not necessarily limited to a square block, and may be configured in a polygonal shape having three or more vertices.

In addition, in the present specification, pixels or pixels are collectively referred to as samples. In addition, using a sample may mean using a pixel value or a pixel value.

1 shows an example of an image coding system according to an embodiment of the present specification. The image coding system may include a source device 10 and a reception device 20. The source device 10 may transmit the encoded video/video information or data in a file or streaming format to the receiving device 20 through a digital storage medium or a network.

The source device 10 may include a video source 11, an encoding device 12, and a transmitter 13. The receiving device 20 may include a receiver 21, a decoding device 22 and a renderer 23. The encoding device 12 may be referred to as a video/image encoding device, and the decoding device 22 may be referred to as a video/image decoding device. The transmitter 13 may be included in the encoding device 12. The receiver 21 may be included in the decoding device 22. The renderer 23 may include a display unit, and the display unit may be configured as a separate device or an external component.

The video source 11 may acquire a video/image through a process of capturing, synthesizing, or generating a video/image. The video source 11 may include a video/image capturing device and/or a video/image generating device. The video/image capturing device may include, for example, one or more cameras, and a video/image archive including previously captured video/images. Video/image generating devices may include, for example, computers, tablets, and smartphones, and may (electronically) generate video/images. For example, a virtual video/image may be generated through a computer, and in this case, a video/image capturing process may be substituted as a process of generating related data.

The encoding device 12 may encode an input video/video. The encoding apparatus 12 may perform a series of procedures such as prediction, transformation, and quantization for compression and coding efficiency. The encoded data (encoded video/video information) may be output in the form of a bitstream.

The transmitter 13 may transmit the encoded video/video information or data output in the form of a bitstream to the receiver 21 of the receiving device 20 through a digital storage medium or a network in a file or streaming form. Digital storage media include USB (universal serial bus), SD card (secure digital card), CD (compact disc), DVD (digital versatile disc), Blu-ray disc, HDD (hard disk drive), SSD (solid state drive) may include a variety of storage media. The transmitter 13 may include an element for generating a media file through a predetermined file format, and may include an element for transmission through a broadcast/communication network. The receiver 21 may extract the bitstream and transmit it to the decoding device 22.

The decoding device 22 may decode the video/video by performing a series of procedures such as inverse quantization, inverse transformation, and prediction corresponding to the operation of the encoding device 12.

The renderer 23 may render the decoded video/image. The rendered video/image may be displayed through the display unit.

2 shows an example of a schematic block diagram of an encoding apparatus in which encoding of a video signal is performed according to an embodiment of the present specification. The encoding device 100 of FIG. 2 may correspond to the encoding device 12 of FIG. 1.

Referring to FIG. 2, the encoding apparatus 100 includes an image partitioning module 110, a subtraction module 115, a transform module 120, and a quantization module. (130), a de-quantization module (140), an inverse-transform module (150), an addition module (155), a filtering module (160), a memory A (memory) 170, an inter prediction module 180, an intra prediction module 185, and an entropy encoding module 190 may be included. The inter prediction unit 180 and the intra prediction unit 185 may be collectively referred to as a prediction unit. That is, the prediction unit may include an inter prediction unit 180 and an intra prediction unit 185. The transform unit 120, the quantization unit 130, the inverse quantization unit 140, and the inverse transform unit 150 may be included in a residual processing unit. The residual processing unit may further include a subtraction unit 115. The above-described image segmentation unit 110, subtraction unit 115, transform unit 120, quantization unit 130, inverse quantization unit 140, inverse transform unit 150, addition unit 155, filtering unit 160 ), the inter prediction unit 180, the intra prediction unit 185, and the entropy encoding unit 190 may be configured by one hardware component (eg, an encoder or a processor) according to an embodiment. In addition, the memory 170 may include a decoded picture buffer (DPB) 175 and may be configured by a digital storage medium.

The image segmentation unit 110 may divide an input image (or picture, frame) input to the encoding apparatus 100 into one or more processing units. For example, the processing unit may be referred to as a coding unit (CU). In this case, the coding unit may be recursively partitioned from a coding tree unit (CTU) or a largest coding unit (LCU) according to a quad-tree binary-tree (QTBT) structure. For example, one coding unit may be divided into a plurality of coding units of a deeper depth based on a quad tree structure and/or a binary tree structure. In this case, for example, a quad tree structure may be applied first and a binary tree structure may be applied later. Alternatively, the binary tree structure may be applied first. A coding procedure according to an embodiment of the present specification may be performed based on a final coding unit that is no longer divided. In this case, the maximum coding unit may be directly used as the final coding unit based on coding efficiency according to image characteristics. Further, if necessary, the coding unit is recursively divided into coding units of a lower depth, so that a coding unit having an optimal size may be used as a final coding unit. Here, the coding procedure may include procedures such as prediction, transformation, and restoration described below. As another example, the processing unit may further include a prediction unit (PU) or a transform unit (TU). In this case, the prediction unit and the transform unit may be divided from the above-described coding units, respectively. The prediction unit may be a unit of sample prediction, and the transform unit may be a unit for inducing a transform coefficient or a unit for inducing a residual signal from the transform coefficient.

The term "unit" used in this document may be used interchangeably with terms such as "block" or "area" in some cases. In this document, the MxN block may represent a set of samples or transform coefficients consisting of M columns and N rows. A sample may generally represent a pixel or a value of a pixel, may represent a pixel/pixel value of a luminance component, or a pixel/pixel value of a saturation component. A sample may be used as a term corresponding to one picture (or image) as a pixel or pel.

The encoding apparatus 100 subtracts a prediction signal (predicted block, prediction sample array) output from the inter prediction unit 180 or the intra prediction unit 185 from the input video signal (original block, original sample array) A signal (residual signal, residual block, residual sample array) may be generated. The generated residual signal is transmitted to the conversion unit 120. In this case, as illustrated, a unit that subtracts the prediction signal (prediction block, prediction sample array) from the input image signal (original block, original sample array) in the encoding apparatus 100 may be referred to as a subtraction unit 115. . The prediction unit may perform prediction on a block to be processed (hereinafter, referred to as a current block) and generate a predicted block including prediction samples for the current block. The prediction module may determine whether intra prediction or inter prediction is applied on a per CU basis. The prediction unit may generate information about prediction, such as prediction mode information, as described later in the description of each prediction mode, and may transmit information about prediction to the entropy encoding unit 190. Information about prediction is encoded by the entropy encoding unit 190 and may be output in the form of a bitstream.

The intra prediction unit 185 may predict the current block by referring to samples in the current picture. The referenced samples may be located in the vicinity of the current block or may be located away from each other according to the prediction mode. In intra prediction, prediction modes may include a plurality of non-directional modes and a plurality of directional modes. The non-directional mode may include, for example, a DC mode and a planar mode (Planar mode). The directional mode may include, for example, 33 directional prediction modes or 65 directional prediction modes according to a detailed degree of the prediction direction. However, this is an example, and more or less directional prediction modes may be used depending on the setting. The intra prediction unit 185 may determine a prediction mode applied to the current block by using the prediction mode applied to the neighboring block.

The inter prediction unit 180 may derive a predicted block for the current block based on a reference block (reference sample array) specified by a motion vector on the reference picture. In this case, in order to reduce the amount of motion information transmitted in the inter prediction mode, the inter prediction unit 180 may predict motion information in units of blocks, subblocks, or samples based on the correlation between motion information between neighboring blocks and the current block have. The motion information may include a motion vector and a reference picture index. The motion information may further include inter prediction direction (L0 prediction, L1 prediction, Bi prediction, etc.) information. In the case of inter prediction, the neighboring block may include a spatial neighboring block existing in the current picture and a temporal neighboring block existing in the reference picture. The reference picture including the reference block and the reference picture including the temporal neighboring block may be the same or different. The temporal neighboring block may be referred to as a collocated reference block or a colCU (colCU), and a reference picture including the temporal neighboring block may be referred to as a collocated picture (colPic). For example, the inter prediction unit 180 constructs a motion information candidate list based on motion information of neighboring blocks, and indicates which candidate is used to derive a motion vector and/or a reference picture index of the current block. Can generate information. Inter prediction may be performed based on various prediction modes. For example, when a skip mode and a merge mode are used, the inter prediction unit 180 may use motion information of a neighboring block as motion information of a current block. In the case of the skip mode, unlike the merge mode, a residual signal is not transmitted. In the case of motion vector prediction (MVP) mode, a motion vector of a neighboring block is used as a motion vector predictor and a motion vector difference (MVD) is signaled to move the current block. Vector can be indicated.

The prediction unit may generate a prediction signal (prediction sample) based on various prediction methods described later. For example, the prediction unit may not only apply intra prediction or inter prediction to predict one block, but also apply intra prediction and inter prediction together (simultaneously). This may be referred to as CIIP (combined inter and intra prediction). Also, the prediction unit may perform intra block copy (IBC) to predict a block. IBC may be used for content (eg, game) video/video coding, such as, for example, screen content coding (SCC). Also, IBC may be referred to as CPR (current picture referencing). IBC basically performs prediction in the current picture, but can be performed similarly to inter prediction in that it derives a reference block in the current picture. That is, the IBC may use at least one of the inter prediction techniques described in this document.

The prediction signal generated by the prediction unit (including the inter prediction unit 180 and/or the intra prediction unit 185) may be used to generate a reconstructed signal or may be used to generate a residual signal. The transform unit 120 may generate transform coefficients by applying a transform technique to the residual signal. For example, the transformation technique uses at least one of DCT (Discrete Cosine Transform), DST (Discrete Sine Transform), KLT (Karhunen-Loeve Transform), GBT (Graph-Based Transform), or CNT (Conditionally Non-linear Transform). Can include. Here, GBT refers to transformation obtained from a graph representing relationship information between pixels. CNT refers to a transformation obtained based on the prediction signal and generating a prediction signal using all previously reconstructed pixels. In addition, the conversion process may be applied to a pixel block having the same size of a square, or may be applied to a block that is not a square or has a variable size.

The quantization unit 130 quantizes the transform coefficients and transmits the quantized transform coefficients to the entropy encoding unit 190. The entropy encoding unit 190 may encode a quantized signal (information on quantized transform coefficients) and output it as a bitstream. Information about the quantized transform coefficients may be referred to as residual information. The quantization unit 130 may rearrange the quantized transform coefficients in a block form into a one-dimensional vector form based on a coefficient scan order, and quantize the quantized transform coefficients based on the characteristics of the quantized transform coefficients in a one-dimensional vector form. It is also possible to generate information about transform coefficients. The entropy encoding unit 190 may perform various encoding techniques such as exponential Golomb, context-adaptive variable length coding (CAVLC), and context-adaptive binary arithmetic coding (CABAC). The entropy encoding unit 190 may encode information necessary for video/image restoration (eg, values of syntax elements) in addition to quantized transform coefficients together or separately. The encoded information (eg, video/video information) may be transmitted or stored in a bitstream format in units of network abstraction layer (NAL) units. The video/video information may further include information on various parameter sets, such as adaptation parameter set (APS), picture parameter set (PPS), sequence parameter set (SPS), or video parameter set (VPS). Signaled/transmitted information and/or syntax elements described later in this document may be encoded through the above-described encoding procedure and included in the bitstream. The bitstream may be transmitted over a network or may be stored in a digital storage medium. Here, the network may include a broadcasting network and/or a communication network, and the digital storage medium may include a storage medium such as USB, SD, CD, DVD, Blu-ray, HDD, and SSD. For the signal output from the entropy encoding unit 190, a transmission unit (not shown) for transmitting and/or a storage unit (not shown) for storing may be configured as internal/external elements of the encoding apparatus 100, or the transmission unit It may be a component of the entropy encoding unit 190.

The quantized transform coefficients output from the quantization unit 130 may be used to generate a reconstructed signal. For example, a residual signal may be restored by applying inverse quantization and inverse transform through the inverse quantization unit 140 and the inverse transform unit 150 in the loop for the quantized transform coefficients. The addition unit 155 adds the reconstructed residual signal to the prediction signal output from the inter prediction unit 180 or the intra prediction unit 185 to obtain a reconstructed signal (a reconstructed picture, a reconstructed block, and a reconstructed sample array). Can be generated. When there is no residual signal for the block to be processed, such as when the skip mode is applied, the predicted block may be used as a reconstructed block. The addition unit 155 may be referred to as a restoration unit or a restoration block generation unit. The generated reconstructed signal may be used for intra prediction of the next processing target block in the current picture, and may be used for inter prediction of the next picture through filtering as described later.

The filtering unit 160 may improve subjective/objective image quality by applying filtering to the reconstructed signal. For example, the filtering unit 160 may apply various filtering methods to the reconstructed picture to generate a modified reconstructed picture, and may transmit the modified reconstructed picture to the DPB 175 of the memory 170. . Various filtering methods may include, for example, deblocking filtering, sample adaptive offset (SAO), adaptive loop filter (ALF), and bilateral filter. The filtering unit 160 may generate filtering information and transmit the filtering information to the entropy encoding unit 190 as described later in the description of each filtering method. The filtering information may be output in the form of a bitstream through entropy encoding in the entropy encoding unit 190.

The modified reconstructed picture transmitted to the DPB 175 may be used as a reference picture in the inter prediction unit 180. When inter prediction is applied, the encoding apparatus 100 may avoid prediction mismatch between the encoding apparatus 100 and the decoding apparatus 200 by using the modified reconstructed picture, and may improve encoding efficiency. The DPB 175 may store the modified reconstructed picture for use as a reference picture in the inter prediction unit 180. The stored motion information may be transmitted to the inter prediction unit 180 for use as motion information of a spatial neighboring block or motion information of a temporal neighboring block. The memory 170 may store reconstructed samples of reconstructed blocks in the current picture, and transfer information on the reconstructed samples to the intra prediction unit 185.

3 shows an example of a schematic block diagram of a decoding apparatus for decoding an image signal according to an embodiment of the present specification. The decoding device 200 of FIG. 3 may correspond to the decoding device 22 of FIG. 1.

Referring to FIG. 3, the decoding apparatus 200 includes an entropy decoding module 210, a de-quantization module 220, an inverse transform module 230, and an adder. (addition module) 235, filtering module 240, memory 250, inter prediction module 260, and intra prediction module 265 may be included. have. The inter prediction unit 260 and the intra prediction unit 265 may be collectively referred to as a prediction module. That is, the prediction unit may include an inter prediction unit 180 and an intra prediction unit 185. The inverse quantization unit 220 and the inverse transform unit 230 may be collectively referred to as a residual processing module. That is, the residual processing unit may include an inverse quantization unit 220 and an inverse transform unit 230. The entropy decoding unit 210, the inverse quantization unit 220, the inverse transform unit 230, the addition unit 235, the filtering unit 240, the inter prediction unit 260, and the intra prediction unit 265 are implemented. It may be configured by one hardware component (eg, a decoder or a processor) according to an example. Also, the memory 250 may include the DPB 255, and may be configured by one hardware component (eg, a memory or a digital storage medium) according to an embodiment.

When a bitstream including video/image information is input, the decoding apparatus 200 may reconstruct an image in response to a process in which the video/image information is processed by the encoding apparatus 100 of FIG. 2. For example, the decoding apparatus 200 may perform decoding using a processing unit applied by the encoding apparatus 100. Thus, upon decoding, the processing unit may be, for example, a coding unit, and the coding unit may be divided from a coding tree unit or a maximum coding unit according to a quad tree structure and/or a binary tree structure. In addition, the reconstructed image signal decoded and output through the decoding device 200 may be reproduced through the playback device.

The decoding apparatus 200 may receive a signal output from the encoding apparatus 100 of FIG. 2 in the form of a bitstream, and the received signal may be decoded through the entropy decoding unit 210. For example, the entropy decoding unit 210 may parse the bitstream to derive information (eg, video/video information) necessary for image restoration (or picture restoration). The video/video information may further include information on various parameter sets, such as adaptation parameter set (APS), picture parameter set (PPS), sequence parameter set (SPS), or video parameter set (VPS). The decoding apparatus may decode a picture based on information on a parameter set. Signaled/received information and/or syntax elements described later in this document may be decoded through a decoding procedure and obtained from a bitstream. For example, the entropy decoding unit 210 acquires information in the bitstream using a coding technique such as exponential Golomb coding, CAVLC, or CABAC, and a value of a syntax element required for image restoration, and a quantized value of a transform coefficient for a residual. Can be printed. In more detail, in the CABAC entropy decoding method, a bin corresponding to each syntax element is received in a bitstream, and information about the syntax element to be decoded and decoding information of a block to be decoded and a neighbor or a symbol/bin decoded in a previous step The symbol corresponding to the value of each syntax element is determined by determining the context model using the information of, and performing arithmetic decoding of the bin by predicting the probability of occurrence of the bin according to the determined context model. Can be generated. In this case, the CABAC entropy decoding method may update the context model using information of the decoded symbol/bin for the context model of the next symbol/bin after the context model is determined. Among the information decoded by the entropy decoding unit 210, information on prediction is provided to the prediction unit (inter prediction unit 260 and intra prediction unit 265), and the register on which entropy decoding is performed by the entropy decoding unit 210 Dual values, that is, quantized transform coefficients and related parameter information may be input to the inverse quantization unit 220. In addition, information about filtering among information decoded by the entropy decoding unit 210 may be provided to the filtering unit 240. Meanwhile, a receiving unit (not shown) for receiving a signal output from the encoding device 100 may be further configured as an inner/outer element of the decoding device 200, or the receiving unit may be a component of the entropy decoding unit 210. May be. Meanwhile, the decoding apparatus 200 according to the present specification may be referred to as a video/video/picture decoding apparatus. The decoding apparatus 200 may be divided into an information decoder (video/video/picture information decoder) and a sample decoder (video/video/picture sample decoder). The information decoder may include an entropy decoding unit 210, and the sample decoder is an inverse quantization unit 220, an inverse transform unit 230, an addition unit 235, a filtering unit 240, a memory 250, and an inter prediction. It may include at least one of the unit 260 and the intra prediction unit 265.

The inverse quantization unit 220 may output transform coefficients through inverse quantization of the quantized transform coefficients. The inverse quantization unit 220 may rearrange the quantized transform coefficients into a two-dimensional block shape. In this case, the reordering may be performed based on the coefficient scan order performed by the encoding apparatus 100. The inverse quantization unit 220 may perform inverse quantization on quantized transform coefficients by using a quantization parameter (eg, quantization step size information) and obtain transform coefficients.

The inverse transform unit 230 inversely transforms the transform coefficients to obtain a residual signal (residual block, residual sample array).

The prediction unit may perform prediction on the current block and generate a predicted block including prediction samples for the current block. The prediction unit may determine whether intra prediction or inter prediction is applied to the current block based on information on prediction output from the entropy decoding unit 210, and may determine a specific intra/inter prediction mode.

The intra prediction unit 265 may predict the current block by referring to samples in the current picture. The referenced samples may be located near the current block or may be spaced apart according to the prediction mode. In intra prediction, prediction modes may include a plurality of non-directional modes and a plurality of directional modes. The intra prediction unit 265 may determine a prediction mode applied to the current block by using the prediction mode applied to the neighboring block.

The inter prediction unit 260 may derive a predicted block for the current block based on a reference block (reference sample array) specified by a motion vector on the reference picture. In this case, in order to reduce the amount of motion information transmitted in the inter prediction mode, motion information may be predicted in units of blocks, subblocks, or samples based on correlation between motion information between neighboring blocks and current blocks. The motion information may include a motion vector and a reference picture index. The motion information may further include inter prediction direction (L0 prediction, L1 prediction, Bi prediction) information. In the case of inter prediction, the neighboring block may include a spatial neighboring block existing in the current picture and a temporal neighboring block existing in the reference picture. For example, the inter prediction unit 260 may construct a motion information candidate list based on neighboring blocks, and derive a motion vector and/or a reference picture index of the current block based on the received candidate selection information. Inter prediction may be performed based on various prediction modes, and information on prediction may include information indicating a mode of inter prediction for a current block.

The addition unit 235 is reconstructed by adding the obtained residual signal to the prediction signal (predicted block, prediction sample array) output from the prediction unit (including the inter prediction unit 260 and/or the intra prediction unit 265). Signals (restored pictures, reconstructed blocks, reconstructed sample arrays) can be generated. When there is no residual for a block to be processed, such as when the skip mode is applied, the predicted block may be used as a reconstructed block.

The addition unit 235 may be referred to as a restoration unit or a restoration block generation unit. The generated reconstructed signal may be used for intra prediction of the next processing target block in the current picture, and may be used for inter prediction of the next picture through filtering as described later.

The filtering unit 240 may improve subjective/objective image quality by applying filtering to the reconstructed signal. For example, the filtering unit 240 may apply various filtering methods to the reconstructed picture to generate a modified reconstructed picture, and may transmit the modified reconstructed picture to the DPB 255 of the memory 250. . Various filtering methods may include, for example, deblocking filtering, sample adaptive offset, adaptive loop filter, and bilateral filter.

The modified reconstructed picture delivered to the DPB 255 of the memory 250 may be used as a reference picture by the inter prediction unit 260. The memory 250 may store motion information of a block from which motion information in a current picture is derived (or decoded) and/or motion information of blocks in a picture that have already been reconstructed. The stored motion information may be transmitted to the inter prediction unit 260 to be used as motion information of a spatial neighboring block or motion information of a temporal neighboring block. The memory 250 may store reconstructed samples of reconstructed blocks in the current picture, and may be transmitted to the intra prediction unit 265.

In the present specification, embodiments described in the filtering unit 160, the inter prediction unit 180, and the intra prediction unit 185 of the encoding apparatus 100 are respectively the filtering unit 240 and the inter prediction unit 260 of the decoding apparatus. ) And the intra prediction unit 265 may be applied to be the same or correspond to each other.

4 shows an example of a content streaming system according to an embodiment of the present specification. Content streaming systems to which the embodiments of the present specification are applied are largely an encoding server 410, a streaming server 420, a web server 430, and a media storage 440. ), a user equipment 450, and a multimedia input device 460.

The encoding server 410 generates a bitstream by compressing content input from a multimedia input device 460 such as a smartphone, a camera, or a camcorder into digital data, and transmits the generated bitstream to the streaming server 420. As another example, when the multimedia input device 460 such as a smartphone, a camera, or a camcorder directly generates a bitstream, the encoding server 410 may be omitted.

The bitstream may be generated by an encoding method or a bitstream generation method to which an embodiment of the present specification is applied, and the streaming server 420 may temporarily store the bitstream while transmitting or receiving the bitstream.

The streaming server 420 transmits multimedia data to the user device 450 based on a user request through the web server 430, and the web server 430 serves as an intermediary that informs the user of what kind of service exists. When a user requests a desired service from the web server 430, the web server 430 transmits information on the requested service to the streaming server 420, and the streaming server 420 transmits multimedia data to the user. In this case, the content streaming system may include a separate control server, and in this case, the control server serves to control commands/responses between devices in the content streaming system.

The streaming server 420 may receive content from the media storage 440 and/or the encoding server 410. For example, when content is received from the encoding server 410, the content may be received in real time. In this case, in order to provide a smooth streaming service, the streaming server 420 may store the bitstream for a predetermined time.

The user device 450 includes, for example, a mobile phone, a smart phone, a laptop computer, a digital broadcasting terminal, a personal digital assistants (PDA), a portable multimedia player (PMP), a navigation system, and a slate PC ( slate PC), tablet PC, ultrabook, wearable device, e.g., smartwatch, smart glass, head mounted display (HMD)), It can include digital TV, desktop computer, and digital signage.

Each server in the content streaming system may be operated as a distributed server, and in this case, data received from each server may be distributedly processed.

5 shows an example of a video signal processing apparatus according to an embodiment of the present specification. The video signal processing apparatus of FIG. 5 may correspond to the encoding apparatus 100 of FIG. 1 or the decoding apparatus 200 of FIG. 2.

The video signal processing apparatus 500 for processing a video signal includes a memory 520 for storing a video signal, and a processor 510 for processing a video signal while being combined with the memory 520. The processor 510 according to the embodiment of the present specification may be configured with at least one processing circuit for processing a video signal, and may process a video signal by executing instructions for encoding/decoding a video signal. That is, the processor 510 may encode original video data or decode an encoded video signal by executing encoding/decoding methods described below. The processor 510 may be composed of one or more processors corresponding to each of the modules of FIG. 2 or 3. The memory 520 may correspond to the memory 170 of FIG. 2 or the memory 250 of FIG. 3.

Partitioning structure

The video/image coding method according to the present specification may be performed based on a split structure described later. Procedures such as prediction, residual processing (e.g., (inverse) transformation, (inverse) quantization), syntax element coding, and filtering, which will be described later in detail, are CTU (coding tree unit) derived based on the load structure, CU (and/ Alternatively, it may be performed based on TU, PU). The block division procedure may be performed by the video division unit 110 of the encoding apparatus 100 described above, and division-related information is (encoded) processed by the entropy encoding unit 190 and transferred to the decoding apparatus 200 in the form of a bitstream. Can be delivered. The entropy decoding unit 210 of the decoding apparatus 200 derives the block division structure of the current block based on the division-related information obtained from the bitstream, and based on this, a series of procedures (e.g., prediction, registration) for decoding an image. Dual processing, block/picture restoration, and in-loop filtering) can be performed.

In video/image coding according to the embodiment of the present specification, an image processing unit may have a hierarchical structure. One picture may be divided into one or more tiles or tile groups. One tile group may include one or more tiles. One tile may contain more than one CTU. The CTU can be divided into one or more CUs. A tile is a rectangular region of CTUs within a particular tile column and a particular tile row in a picture. The tile group may include an integer number of tiles according to a tile raster scan in a picture. The tile group header may convey information/parameters applicable to the corresponding tile group. When the encoding device 100/decoding device 200 has a multi-core processor, an encoding/decoding procedure for a tile or a group of tiles may be processed in parallel. Here, the tile group may have one type of tile groups including an intra (I) tile group, a predictive (P) tile group, and a bi-predictive (B) tile group. For prediction of blocks in an I tile group, inter prediction is not used and only intra prediction can be used. Of course, even for the I tile group, a coded original sample value may be signaled without prediction. Intra prediction or inter prediction may be used for blocks in a P tile group, and when inter prediction is used, only uni prediction may be used. Meanwhile, intra prediction or inter prediction may be used for blocks in the B tile group, and when inter prediction is used, not only unidirectional prediction but also bi prediction may be used.

6 illustrates an example of a picture division structure according to an embodiment of the present specification. In FIG. 6, a picture having 216 (18 by 12) luminance CTUs is divided into 12 tiles and 3 tile groups.

The encoder determines the size of a tile/tile group and a maximum and minimum coding unit according to a characteristic (e.g., resolution) of a video image or in consideration of coding efficiency or parallel processing, and provides information about this or information for inducing it. It can be included in the bitstream.

The decoder may obtain information indicating whether the tile/tile group of the current picture and the CTU in the tile are divided into a plurality of coding units. Coding efficiency can be increased if such information is not always acquired (decoded) by the decoder, but is acquired (decoded) only under certain conditions.

The tile group header (tile group header syntax) may include information/parameters commonly applicable to the tile group. APS (APS syntax) or PPS (PPS syntax) may include information/parameters that can be commonly applied to one or more pictures. SPS (SPS syntax) may include information/parameters that can be commonly applied to one or more sequences. VPS (VPS syntax) may include information/parameters that can be commonly applied to the entire video. The high-level syntax in the present specification may include at least one of APS syntax, PPS syntax, SPS syntax, and VPS syntax.

Further, for example, information on the division and configuration of a tile/tile group may be configured in an encoder through a higher level syntax and then transmitted to a decoder in the form of a bitstream.

7A to 7D illustrate examples of a block division structure according to an embodiment of the present specification. 7A is a QT (quadtree, QT), FIG. 7b is a binary tree (BT), and FIG. 7c is a ternary tree (TT) and FIG. 7d shows an example of block division structures by an asymmetric tree (AT). do.

In a video coding system, one block may be divided based on a QT division scheme. In addition, one subblock divided by the QT division method may be further divided recursively according to the QT division method. A leaf block that is no longer divided by the QT division method may be divided by at least one of BT, TT, or AT. BT can have two types of division, such as horizontal BT (2NxN, 2NxN) and vertical BT (Nx2N, Nx2N). TT may have two types of division, such as horizontal TT (2Nx1/2N, 2NxN, 2Nx1/2N) and vertical TT (1/2Nx2N, Nx2N, 1/2Nx2N). AT is horizontal-up AT (2Nx1/2N, 2Nx3/2N), horizontal-down AT (2Nx3/2N, 2Nx1/2N), vertical-left AT ( It can have four types of division: 1/2Nx2N, 3/2Nx2N), and vertical-right AT (3/2Nx2N, 1/2Nx2N). Each BT, TT, AT can be further divided recursively using BT, TT, AT.

7A shows an example of QT division. Block A may be divided into four sub-blocks (A0, A1, A2, A3) by QT. Sub-block A1 may be divided into four sub-blocks (B0, B1, B2, B3) by QT again.

7B shows an example of BT segmentation. Block B3 that is no longer divided by QT may be divided by vertical BT (C0, C1) or horizontal BT (D0, D1). Like block C0, each sub-block may be further divided recursively in the form of horizontal BT (E0, E1) or vertical BT (F0, F1).

7C shows an example of TT partitioning. Block B3 which is no longer divided by QT may be divided into vertical TT (C0, C1, C2) or horizontal TT (D0, D1, D2). Like block C1, each sub-block may be further divided recursively in the form of horizontal TT (E0, E1, E2) or vertical TT (F0, F1, F2).

7D shows an example of AT partitioning. Block B3, which is no longer divided by QT, can be divided into vertical ATs (C0, C1) or horizontal ATs (D0, D1). Like block C1, each sub-block can be further divided recursively in the form of a horizontal AT (E0, E1) or a vertical TT (F0, F1).

Meanwhile, BT, TT, and AT division can be applied together in one block. For example, a sub-block divided by BT may be divided by TT or AT. In addition, sub-blocks divided by TT may be divided by BT or AT. Sub-blocks divided by AT may be divided by BT or TT. For example, after horizontal BT division, each sub-block may be divided by vertical BT. In addition, after vertical BT division, each sub-block may be divided by horizontal BT. In this case, the order of division is different, but the shape of the final division is the same.

In addition, when the block is divided, the order of searching for the block may be variously defined. In general, a search is performed from left to right and from top to bottom, and searching for a block means the order of determining whether to divide additional blocks of each divided sub-block, or if the block is no longer divided, each sub It may mean an encoding order of a block, or a search order when a subblock refers to information of another neighboring block.

In addition, virtual pipeline data units (VPDUs) may be defined for intra-picture pipeline processing. VPDUs may be defined as non-overlapping units within one picture. In a hardware decoder, successive VPDUs can be processed simultaneously by multiple pipeline stages. The VPDU size is roughly proportional to the buffer size in most pipeline stages. Therefore, keeping the VDPU size small is important when considering the buffer size from a hardware perspective. In most hardware decoders, the VPDU size can be set equal to the maximum TB size. For example, the VPDU size may be 64x64 (64x64 luminance samples) size. However, this is an example, and the VPDU size may be changed (increased or decreased) in consideration of the TT and/or BT partition described above.

8 shows an example of a case in which TT and BT division are restricted according to an embodiment of the present specification. In order to maintain the VPDU size at 64x64 luminance samples size, at least one of the following restrictions may be applied as illustrated in FIG. 8.

-TT split is not allowed for a CU with either width or height, or both width and height equal to the width or height, or for a CU with both width and height equal to 128. 128).

-For a 128xN CU with N <= 64 (ie width equal to 128 and height smaller than) 128xN (N <= 64) (i.e., width 128 and height less than 128) CU 128), horizontal BT is not allowed).

-For an Nx128 CU with N <= 64 (ie height equal to 128 and width smaller than) Nx128 (N <= 64) (i.e., the height is 128 and the width is less than 128) 128), vertical BT is not allowed).

Video/video coding procedure

In video/video coding, pictures constituting the video/video may be encoded/decoded according to a series of decoding orders. A picture order corresponding to an output order of a decoded picture may be set differently from a decoding order, and based on this, not only forward prediction but also backward prediction may be performed during inter prediction.

9 shows an example of a flowchart for encoding a picture constituting a video signal according to an embodiment of the present specification. In FIG. 9, step S910 may be performed by the

prediction units

180 and 185 of the encoding apparatus 100 described in FIG. 2, and step S920 may be performed by the

residual processing units

115, 120, and 130. , S930 may be performed by the entropy encoding unit 190. Step S910 may include an inter/intra prediction procedure described in this document, step S920 may include a residual processing procedure described in this document, and step S930 includes an information encoding procedure described in this document. can do.

Referring to FIG. 9, the picture encoding procedure is not only a procedure of encoding information for picture restoration (eg, prediction information, residual information, partitioning information) schematically as described in FIG. 2 to output in a bitstream format, A procedure for generating a reconstructed picture for the current picture and a procedure for applying in-loop filtering to the reconstructed picture (optional) may be included. The encoding apparatus 100 may derive (modified) residual samples from the quantized transform coefficients through the inverse quantization unit 140 and the inverse transform unit 150, and prediction samples corresponding to the output of step S910 and ( A reconstructed picture may be generated based on the modified) residual samples. The reconstructed picture generated in this way may be the same as the reconstructed picture generated by the decoding apparatus 200 described above. A modified reconstructed picture can be generated through an in-loop filtering procedure for the reconstructed picture, which can be stored in the memory 170 (DPB 175), and, as in the case of the decoding device 200, a subsequent picture It can be used as a reference picture in an inter prediction procedure upon encoding of. As described above, in some cases, some or all of the in-loop filtering procedure may be omitted. When the in-loop filtering procedure is performed, (in-loop) filtering-related information (parameters) may be encoded by the entropy encoding unit 190 and output in the form of a bitstream, and the decoding apparatus 200 The in-loop filtering procedure may be performed in the same manner as the encoding apparatus 100.

Through this in-loop filtering procedure, noise generated during video/video coding such as blocking artifacts and ringing artifacts can be reduced, and subjective/objective visual quality can be improved. In addition, by performing the in-loop filtering procedure in both the encoding device 100 and the decoding device 200, the encoding device 100 and the decoding device 200 can derive the same prediction result, and increase the reliability of picture coding. , It is possible to reduce the amount of data transmitted for picture coding.

10 shows an example of a flowchart for decoding a picture constituting a video signal according to an embodiment of the present specification. Step S1010 may be performed by the entropy decoding unit 210 of the decoding apparatus 200 of FIG. 3, step S1020 may be performed by the

prediction units

260 and 265, and step S1030 may be performed by the residual processing unit ( 220, 230), step S1040 may be performed by the addition unit 235, step S1050 may be performed by the filtering unit 240. Step S1010 may include the information decoding procedure described in this document, step S1020 may include the inter/intra prediction procedure described in this document, and step S1030 includes the residual processing procedure described in this document. In addition, step S1040 may include the block/picture restoration procedure described in this document, and step S1050 may include the in-loop filtering procedure described in this document.

Referring to FIG. 10, the picture decoding procedure is schematically a procedure for obtaining image/video information (through decoding) from a bitstream (S1010), a picture restoration procedure (S1020 to S1040), and a reconstructed picture as described in FIG. It may include an in-loop filtering procedure for (S1050). The picture restoration procedure is based on prediction samples and residual samples obtained through the process of inter/intra prediction (S1020) and residual processing (S1030, inverse quantization and inverse transformation of a quantized code or coefficient) described in this document. Can be done. A modified reconstructed picture may be generated through an in-loop filtering procedure for a reconstructed picture generated through a picture restoration procedure, and the modified reconstructed picture may be output as a decoded picture, and the decoding apparatus 200 It is stored in the DPB 255 of and can be used as a reference picture in inter prediction train when decoding a picture later. In some cases, the in-loop filtering procedure may be omitted, and in this case, the reconstructed picture may be output as a decoded picture, stored in the DPB 255 of the decoding device 200, and referenced in the inter prediction train when decoding a subsequent picture. Can be used as a picture. The in-loop filtering procedure (S1050) may include a deblocking filtering procedure, a sample adaptive offset (SAO) procedure, an adaptive loop filter (ALF) procedure, and/or a bi-lateral filter procedure as described above. And some or all of them may be omitted. In addition, one or some of the deblocking filtering procedure, the SAO procedure, the ALF procedure, and the bilateral filter procedure may be sequentially applied, or all may be sequentially applied. For example, after the deblocking filtering procedure is applied to the reconstructed picture, the SAO procedure may be performed. Also, for example, after the deblocking filtering procedure is applied to the reconstructed picture, the ALF procedure may be performed. This may be similarly performed in the encoding device 100.

As described above, not only the decoding apparatus 200 but also the encoding apparatus 100 may perform a picture restoration procedure. A reconstructed block may be generated based on intra prediction/inter prediction for each block, and a reconstructed picture including the reconstructed blocks may be generated. When the current picture/slice/tile group is an I picture/slice/tile group, blocks included in the current picture/slice/tile group may be reconstructed based only on intra prediction. In this case, inter prediction may be applied to some blocks in the current picture/slice/tile group, and intra prediction may be applied to the remaining some blocks. The color component of a picture may include a luminance component and a chrominance component, and the methods and embodiments proposed in this document may be applied to the luminance component and the chrominance component unless explicitly limited in this document.

Example of coding hierarchy and structure

The coded image is a video coding layer (VCL) that deals with the decoding process of the image and itself, a subsystem that transmits and stores encoded information, and a network abstraction (NAL) that exists between the VCL and the subsystem and is responsible for network adaptation. layer).

VCL data including video data (tile group data) compressed in the VCL is generated, or a parameter set including information such as PPS (picture parameter set), SPS (sequence parameter set), VPS (video parameter set), or An additionally required SEI (supplemental enhancement information) message may be generated in the process of decoding an image.

In NAL, header information (NAL unit data) may be added to a raw byte sequence payload (RBSP) generated in VCL to generate a NAL unit. In this case, the RBSP may refer to tile group data, parameter set, and SEI message generated in the VCL. In the NAL unit header, NAL unit type information specified according to RBSP data included in the corresponding NAL unit may be included.

As shown in FIG. 11, the NAL unit may be divided into a VCL NAL unit and a Non-VCL NAL unit according to an RBSP generated from VCL. The VCL NAL unit may mean a NAL unit that includes information about an image (tile group data), and the Non-VCL NAL unit is an NAL that includes information (parameter set or SEI message) necessary for decoding an image. It can mean a unit.

The above-described VCL NAL unit and Non-VCL NAL unit may be transmitted through a network with header information added according to the data standard of the sub-system. For example, the NAL unit may be converted into a data format of a predetermined standard such as an H.266/VVC file format, a real-time transport protocol (RTP), and a transport stream (TS) and then transmitted through various networks.

As described above, the NAL unit type may be specified according to the RBSP data structure included in the corresponding NAL unit, and information on the NAL unit type may be stored in the NAL unit header and signaled.

For example, the NAL unit may be largely classified into a VCL NAL unit type and a Non-VCL NAL unit type according to whether or not information on an image (tile group data) is included. The VCL NAL unit type may be classified according to the nature and type of a picture included in the VCL NAL unit, and the non-VCL NAL unit type may be classified according to the type of the parameter set.

The following is an example of the NAl unit type specified according to the type of the parameter set included in the Non-VCL NAL unit type.

-APS (Adaptation Parameter Set) NAL unit: A type for a NAL unit including APS

-Video Parameter Set (VPS) NAL unit: A type for a NAL unit including a VPS

-SPS (Sequence Parameter Set) NAL unit: a type for a NAL unit including SPS

-PPS (Picture Parameter Set) NAL unit: A type for a NAL unit including PPS

The above-described NAL unit types have syntax information for the NAL unit type, and the syntax information may be stored in the NAL unit header and signaled. For example, syntax information may be nal_unit_type, and NAL unit types may be specified by nal_unit_type values.

The tile group header (tile group header syntax) may include information/parameters commonly applicable to the tile group. APS (APS syntax) or PPS (PPS syntax) may include information/parameters that can be commonly applied to one or more pictures. SPS (SPS syntax) may include information/parameters that can be commonly applied to one or more sequences. VPS (VPS syntax) may include information/parameters that can be commonly applied to the entire video. In this specification, the higher-level syntax may include at least one of APS syntax, PPS syntax, SPS syntax, and VPS syntax.

In this specification, the image/video information encoded by the encoding device 100 by the decoding device 200 and signaled in the form of a bitstream includes intra-picture partitioning-related information, intra/inter prediction information, residual information, and in-loop filtering information. In addition, information included in the APS, information included in the PPS, information included in the SPS, and/or the information included in the VPS may be included.

Inter prediction

Hereinafter, an inter prediction technique according to an embodiment of the present specification will be described. The inter prediction described below may be performed by the inter prediction unit 180 of the encoding apparatus 100 of FIG. 2 or the inter prediction unit 260 of the decoding apparatus 200 of FIG. 3. In addition, data encoded according to an embodiment of the present specification may be stored in the form of a bitstream.

The prediction unit of the encoding device 100/decoding device 200 may derive a prediction sample by performing inter prediction in block units. Inter prediction may represent prediction derived in a method dependent on data elements (eg, sample values, motion information, etc.) of a picture(s) other than the current picture (Inter prediction can be a prediction derived in a manner). that is de-pendent on data elements (eg, sample values or motion information) of picture(s) other than the current picture). When inter prediction is applied to the current block, a predicted block (prediction sample array) for the current block is derived based on a reference block (reference sample array) specified by a motion vector on a reference picture indicated by a reference picture index. I can. In this case, in order to reduce the amount of motion information transmitted in the inter prediction mode, motion information of the current block may be predicted in units of blocks, subblocks, or samples based on correlation between motion information between neighboring blocks and the current block. The motion information may include a motion vector and a reference picture index. The motion information may further include inter prediction type (L0 prediction, L1 prediction, Bi prediction, etc.) information. When inter prediction is applied, the neighboring block may include a spatial neighboring block existing in the current picture and a temporal neighboring block existing in the reference picture. The reference picture including the reference block and the reference picture including the temporal neighboring block may be the same or different. The temporal neighboring block may be called a collocated reference block, a co-located CU (colCU), or the like, and a reference picture including a temporal neighboring block may be referred to as a collocated picture (colPic). . For example, a motion information candidate list may be constructed based on neighboring blocks of the current block, and a flag indicating which candidate is selected (used) to derive a motion vector and/or a reference picture index of the current block, or Index information may be signaled. Inter prediction may be performed based on various prediction modes. For example, in the case of a skip mode and a merge mode, motion information of a current block may be the same as motion information of a selected neighboring block. In the case of the skip mode, unlike the merge mode, a residual signal may not be transmitted. In the case of a motion vector prediction (MVP) mode, a motion vector of a selected neighboring block is used as a motion vector predictor, and a motion vector difference may be signaled. In this case, the motion vector of the current block may be derived using the sum of the motion vector predictor and the motion vector difference.

12 is an example of a flowchart for inter prediction in a process of encoding a video signal according to an embodiment of the present specification, and FIG. 13 illustrates an example of an inter prediction unit in an encoding apparatus according to an embodiment of the present specification.

The encoding apparatus 100 performs inter prediction on the current block (S1210). The encoding apparatus 100 may derive inter prediction mode and motion information of the current block and generate prediction samples of the current block. Here, the procedure of determining the inter prediction mode, deriving motion information, and generating prediction samples may be performed simultaneously, or one procedure may be performed before the other procedure. For example, the inter prediction unit 180 of the encoding apparatus 100 may include a prediction mode determining unit 181, a motion information deriving unit 182, and a predicted sample deriving unit 183, and the prediction mode determining unit A prediction mode for the current block may be determined at 181, motion information of the current block may be derived by the motion information deriving unit 182, and prediction samples of the current block may be derived by the predicted sample deriving unit 183. For example, the inter prediction unit 180 of the encoding apparatus 100 searches for a block similar to the current block within a predetermined area (search area) of reference pictures through motion estimation, and searches for a block similar to the current block. It is possible to derive a reference block whose difference is less than a minimum or a certain standard. Based on this, a reference picture index indicating a reference picture in which the reference block is located may be derived, and a motion vector may be derived based on a position difference between the reference block and the current block. The encoding apparatus 100 may determine a mode applied to the current block among various prediction modes. The encoding apparatus 100 may compare rate-distortion (RD) costs for various prediction modes and determine an optimal prediction mode for the current block.

For example, when the skip mode or the merge mode is applied to the current block, the encoding apparatus 100 configures a merge candidate list to be described later, and the current block and the middle of the reference blocks indicated by merge candidates included in the merge candidate list. It is possible to derive a reference block whose difference from the current block is less than a minimum or a certain standard. In this case, a merge candidate associated with the derived reference block is selected, and merge index information indicating the selected merge candidate may be generated and signaled to the decoding apparatus 200. Motion information of the current block may be derived using motion information of the selected merge candidate.

As another example, when the (A)MVP mode is applied to the current block, the encoding apparatus 100 configures a (A)MVP candidate list to be described later, and (A)motion vector predictor (MVP) candidates included in the MVP candidate list The motion vector of the selected MVP candidate may be used as the MVP of the current block. In this case, for example, a motion vector indicating a reference block derived by motion estimation described above may be used as a motion vector of the current block, and among MVP candidates, a motion vector having the smallest difference from the motion vector of the current block is selected. The MVP candidate to have becomes the selected MVP candidate. A motion vector difference (MVD), which is a difference obtained by subtracting MVP from the motion vector of the current block, may be derived. In this case, information on the MVD may be signaled to the decoding apparatus 200. In addition, when the (A)MVP mode is applied, the value of the reference picture index may be separately signaled to the decoding apparatus 200 by configuring reference picture index information.

The encoding apparatus 100 may derive residual samples based on the prediction samples (S1220). The encoding apparatus 100 may derive residual samples by comparing the original samples of the current block with the prediction samples.

The encoding apparatus 100 encodes video information including prediction information and residual information (S1230). The encoding apparatus 100 may output the encoded image information in the form of a bitstream. The prediction information is information related to a prediction procedure and may include prediction mode information (eg, skip flag, merge flag, or mode index) and motion information. The motion information may include candidate selection information (eg, merge index, mvp flag, or mvp index) that is information for deriving a motion vector. Further, the motion information may include information on the MVD and/or reference picture index information described above. Further, the motion information may include information indicating whether L0 prediction, L1 prediction, or bi prediction is applied. The residual information is information about residual samples. The residual information may include information about quantized transform coefficients for residual samples. The prediction mode information and motion information may be collectively referred to as inter prediction information.

The output bitstream may be stored in a (digital) storage medium and transmitted to a decoding device, or may be transmitted to a decoding device through a network.

Meanwhile, as described above, the encoding apparatus may generate a reconstructed picture (including reconstructed samples and a reconstructed block) based on the reference samples and the residual samples. This is because the encoding device 100 derives the same prediction result as that performed by the decoding device 200, and coding efficiency can be improved through this. Accordingly, the encoding apparatus 100 may store a reconstructed picture (or reconstructed samples, and a reconstructed block) in a memory and use it as a reference picture for inter prediction. As described above, an in-loop filtering procedure or the like may be further applied to the reconstructed picture.

14 is an example of a flowchart for inter prediction in a process of decoding a video signal according to an embodiment of the present specification, and FIG. 15 shows an example of an inter prediction unit in a decoding apparatus according to an embodiment of the present specification.

The decoding apparatus 200 may perform an operation corresponding to the operation performed by the encoding apparatus 100. The decoding apparatus 200 may perform prediction on the current block and derive prediction samples based on the received prediction information.

In more detail, the decoding apparatus 200 may determine a prediction mode for the current block based on the received prediction information (S1410). The decoding apparatus 200 may determine which inter prediction mode is applied to the current block based on prediction mode information in the prediction information.

For example, the decoding apparatus 200 may determine whether the merge mode is applied to the current block or the (A)MVP mode is determined based on the merge flag. Alternatively, the decoding apparatus 200 may select one of various inter prediction mode candidates based on a mode index. Inter prediction mode candidates may include a skip mode, a merge mode, and/or (A)MVP mode, or may include various inter prediction modes described below.

The decoding apparatus 200 derives motion information of the current block based on the determined inter prediction mode (S1420). For example, when the skip mode or the merge mode is applied to the current block, the decoding apparatus 200 may configure a merge candidate list to be described later, and select one merge candidate from among merge candidates included in the merge candidate list. . The selection of a merge candidate may be performed based on a merge index. Motion information of the current block may be derived from motion information of the selected merge candidate. Motion information of the selected merge candidate may be used as motion information of the current block.

As another example, when the (A)MVP mode is applied to the current block, the decoding apparatus 200 constructs a (A)MVP candidate list to be described later, and (A) a selected MVP candidate among MVP candidates included in the MVP candidate list. The motion vector of can be used as the MVP of the current block. The selection of MVP may be performed based on the above-described selection information (MVP flag or MVP index). In this case, the decoding apparatus 200 may derive the MVD of the current block based on the information on the MVD, and may derive a motion vector of the current block based on the MVP and the MVD of the current block. Also, the decoding apparatus 200 may derive the reference picture index of the current block based on the reference picture index information. The picture indicated by the reference picture index in the reference picture list for the current block may be derived as a reference picture referenced for inter prediction of the current block.

Meanwhile, as described later, motion information of the current block may be derived without configuring a candidate list. In this case, motion information of the current block may be derived according to a procedure disclosed in a prediction mode to be described later. In this case, the configuration of the candidate list as described above may be omitted.

The decoding apparatus 200 may generate prediction samples for the current block based on motion information of the current block (S1430). In this case, the decoding apparatus 200 may derive a reference picture based on the reference picture index of the current block, and may derive the prediction samples of the current block by using samples of the reference block indicated on the reference picture by the motion vector of the current block. . In this case, as will be described later, a prediction sample filtering procedure may be further performed on all or part of the prediction samples of the current block in some cases.

For example, the inter prediction unit 260 of the decoding apparatus 200 may include a prediction mode determination unit 261, a motion information derivation unit 262, and a prediction sample derivation unit 263, and a prediction mode determination unit A prediction mode for the current block is determined based on the prediction mode information received at (181), and motion information (motion vector) of the current block is determined based on the information on the motion information received from the motion information derivation unit 182. And/or a reference picture index), and the prediction sample derivation unit 183 may derive prediction samples of the current block.

The decoding apparatus 200 generates residual samples for the current block based on the received residual information (S1440). The decoding apparatus 200 may generate reconstructed samples for the current block based on the prediction samples and the residual samples, and generate a reconstructed picture based on this (S1450). Thereafter, as described above, an in-loop filtering procedure or the like may be further applied to the reconstructed picture.

As described above, the inter prediction procedure may include determining an inter prediction mode, deriving motion information according to the determined prediction mode, and performing prediction based on the derived motion information (generating a prediction sample).

Inter prediction mode determination

Various inter prediction modes may be used for prediction of a current block in a picture. For example, various modes, such as a merge mode, a skip mode, an MVP mode, and an affine mode, may be used. A decoder side motion vector refinement (DMVR) mode, an adaptive motion vector resolution (AMVR) mode, or the like may be further used as an auxiliary mode. The afine mode may also be referred to as an affine motion prediction mode. The MVP mode may also be called an advanced motion vector prediction (AMVP) mode.

Prediction mode information indicating the inter prediction mode of the current block may be signaled from the encoding device to the decoding device 200. The prediction mode information may be included in the bitstream and received by the decoding apparatus 200. The prediction mode information may include index information indicating one of a plurality of candidate modes. Alternatively, the inter prediction mode may be indicated through hierarchical signaling of flag information. In this case, the prediction mode information may include one or more flags. For example, the encoding apparatus 100 signals the skip flag to indicate whether to apply the skip mode, and when the skip mode is not applied, signals the merge flag to indicate whether to apply the merge mode, and when the merge mode is not applied It may indicate that the MVP mode is applied or a flag for additional classification may be further signaled. The afine mode may be signaled as an independent mode, or may be signaled as a mode dependent on the merge mode or the MVP mode. For example, the affine mode may be composed of one candidate of a merge candidate list or an MVP candidate list, as described later.

Derive motion information

The encoding device 100 or the decoding device 200 may perform inter prediction using motion information of the current block. The encoding apparatus 100 may derive optimal motion information for the current block through a motion estimation procedure. For example, the encoding apparatus 100 may search for a similar reference block with high correlation using the original block in the original picture for the current block in units of fractional pixels within a predetermined search range within the reference picture, and through this Can be derived. The similarity of the block may be derived based on the difference between the phase-based sample values. For example, the similarity of blocks may be calculated based on a sum of absolute difference (SAD) between a current block (or a template of a current block) and a reference block (or a template of a reference block). In this case, motion information may be derived based on the reference block having the smallest SAD in the search area. The derived motion information may be signaled to the decoding apparatus according to various methods based on the inter prediction mode.

Merge mode and skip mode

When the merge mode is applied, motion information of a current prediction block is not directly transmitted, and motion information of a current prediction block is derived using motion information of a neighboring prediction block. Accordingly, the encoding apparatus 100 may indicate motion information of the current prediction block by transmitting flag information indicating that the merge mode has been used and a merge index indicating which prediction block is used.

In order to perform a merge mode, the encoding apparatus 100 must search for a merge candidate block used to induce motion information of a current prediction block. For example, up to five merge candidate blocks may be used, but the present specification is not limited thereto. In addition, the maximum number of merge candidate blocks may be transmitted in a slice header or a tile group header, and the present specification is not limited thereto. After finding the merge candidate blocks, the encoding apparatus 100 may generate a merge candidate list and select a merge candidate block having the lowest cost among them as the final merge candidate block.

The merge candidate list may use, for example, 5 merge candidate blocks. For example, four spatial merge candidates and one temporal merge candidate can be used.

Referring to FIG. 16, for prediction of a current block, a left neighboring block A1, a bottom-left neighboring block A0, a top-right neighboring block B0, and an upper neighboring block B1. ), and at least one of the top-left neighboring block B2 may be used. The merge candidate list for the current block may be configured based on the procedure shown in FIG. 17.

The coding apparatus (encoding apparatus 100 or decoding apparatus 200) searches for spatial neighboring blocks of the current block and inserts the derived spatial merge candidates into the merge candidate list (S1710). For example, the spatial neighboring blocks may include a block around a lower left corner of a current block, a block around a left, a block around an upper right corner, a block around an upper side, and blocks around an upper left corner. However, as an example, in addition to the spatial neighboring blocks described above, additional neighboring blocks such as a right peripheral block, a lower peripheral block, and a right lower peripheral block may be further used as the spatial neighboring blocks. The coding apparatus may detect available blocks by searching spatial neighboring blocks based on priority, and derive motion information of the detected blocks as spatial merge candidates. For example, the encoding device 100 or the decoding device 200 searches the five blocks shown in FIG. 16 in the order of A1, B1, B0, A0, B2, and sequentially indexes the available candidates to obtain a merge candidate. It can be organized as a list.

The coding apparatus inserts a temporal merge candidate derived by searching for a temporal neighboring block of the current block into the merge candidate list (S1720). The temporal neighboring block may be located on a reference picture that is a picture different from the current picture in which the current block is located. A reference picture in which a temporal neighboring block is located may be referred to as a collocated picture or a col picture. The temporal neighboring block may be searched in the order of the lower right corner neighboring block and the lower right center block of the collocated block with respect to the current block on the collocated picture. Meanwhile, when motion data compression is applied, specific motion information may be stored as representative motion information for each predetermined storage unit in a collocated picture. In this case, it is not necessary to store motion information for all blocks in the predetermined storage unit, and motion data compression effect can be obtained through this. In this case, the predetermined storage unit may be predetermined, for example, in a 16x16 sample unit, an 8x8 sample unit, or the like, or size information for a predetermined storage unit may be signaled from the encoding device 100 to the decoding device 200 have. When motion data compression is applied, motion information of a temporal neighboring block may be replaced with representative motion information of a predetermined storage unit in which a temporal neighboring block is located. In other words, in this case, in terms of implementation, it is not a prediction block located at the coordinates of a temporal neighboring block, but an arithmetic right shift by a certain value based on the coordinates of the temporal neighboring block (top left sample position) and then the arithmetic left shifted position is covered. A temporal merge candidate may be derived based on motion information of the prediction block. For example, if the constant storage unit is a 2nx2n sample unit, if the coordinates of the temporal neighboring block are (xTnb, yTnb), the modified positions ((xTnb >> n) << n), (yTnb >> n) The motion information of the prediction block located at << n)) may be used for the temporal merge candidate. Specifically, for example, if the predetermined storage unit is a 16x16 sample unit, if the coordinates of the temporal neighboring block are (xTnb, yTnb), the modified positions ((xTnb >> 4) << 4), (yTnb >> 4) Motion information of the prediction block located at << 4)) may be used for a temporal merge candidate. Or, for example, if the constant storage unit is an 8x8 sample unit, if the coordinates of the temporal neighboring block are (xTnb, yTnb), the modified positions ((xTnb >> 3) << 3), (yTnb >> 3 ) << 3)) motion information of a prediction block may be used for a temporal merge candidate.

The coding apparatus may check whether the number of current merge candidates is smaller than the number of maximum merge candidates (S1730). The maximum number of merge candidates may be predefined or signaled from the encoding device 100 to the decoding device 200. For example, the encoding apparatus 100 may generate information on the number of maximum merge candidates, encode, and transmit the information to the decoding apparatus 200 in the form of a bitstream. When the maximum number of merge candidates is filled, the subsequent candidate addition process may not proceed.

As a result of checking, if the number of current merge candidates is less than the number of the maximum merge candidates, the coding apparatus inserts an additional merge candidate into the merge candidate list (S1740). Additional merge candidates include, for example, adaptive temporal motion vector prediction (ATMVP), combined bi-predictive merge candidate (when the slice type of the current slice is B type), and/or zero vector merge. Can include candidates.

MVP mode

The MVP mode may be referred to as AMVP (advanced MVP or adaptive MVP). When the MVP mode is applied, a motion vector predictor using a motion vector of a reconstructed spatial neighboring block (eg, neighboring block in FIG. 16) and/or a motion vector corresponding to a temporal neighboring block (or Col block) A (motion vector predictor, MVP) candidate list may be generated. That is, a motion vector of the reconstructed spatial neighboring block and/or a motion vector corresponding to the temporal neighboring block may be used as a motion vector predictor candidate. The information on prediction may include selection information (eg, an MVP flag or an MVP index) indicating an optimal motion vector predictor candidate selected from among motion vector predictor candidates included in the list. In this case, the prediction unit may select a motion vector predictor of the current block from among motion vector predictor candidates included in the motion vector candidate list using the selection information. The prediction unit of the encoding apparatus 100 may obtain a motion vector difference (MVD) between the motion vector of the current block and the motion vector predictor, encode the motion vector, and output it in the form of a bitstream. That is, MVD may be obtained by subtracting the motion vector predictor from the motion vector of the current block. In this case, the prediction unit of the decoding apparatus may obtain a motion vector difference included in the prediction information, and derive the motion vector of the current block by adding the motion vector difference and the motion vector predictor. The prediction unit of the decoding apparatus may obtain or derive a reference picture index indicating a reference picture from the prediction information. For example, the motion vector predictor candidate list may be configured as shown in FIG. 18.

Referring to FIG. 18, the coding apparatus searches for a spatial candidate block for motion vector prediction and inserts it into a prediction candidate list (S1810). For example, the coding apparatus may search for neighboring blocks according to a predetermined search order, and add information on neighboring blocks that satisfy a condition for a spatial candidate block to a prediction candidate list (MVP candidate list).

After constructing a spatial candidate block list, the coding apparatus compares the number of spatial candidate lists included in the prediction candidate list with a preset reference number (eg, 2) (S1820). When the number of spatial candidate lists included in the prediction candidate list is greater than or equal to the reference number (eg, 2), the coding apparatus may terminate the construction of the prediction candidate list.

However, when the number of spatial candidate lists included in the prediction candidate list is less than the reference number (eg, 2), the coding apparatus searches for a temporal candidate block and inserts it into the prediction candidate list (S1830), and the temporal candidate block is used. If not possible, a zero motion vector is added to the prediction candidate list (S1840).

A predicted block for the current block may be derived based on motion information derived according to the prediction mode. The predicted block may include predicted samples (prediction sample array) of the current block. When the motion vector of the current block indicates a fractional sample unit, an interpolation procedure may be performed, through which prediction samples of the current block may be derived based on reference samples of the fractional sample unit within a reference picture. . When affine inter prediction is applied to the current block, prediction samples may be generated based on a motion vector per sample/subblock. When bi-direction prediction is applied, the prediction samples derived based on the first direction prediction (eg, L0 prediction) and the prediction samples derived based on the second direction prediction (eg, L1 prediction) Final prediction samples can be derived through weighted summation (according to the phase). As described above, reconstructed samples and reconstructed pictures may be generated based on the derived prediction samples, and then a procedure such as in-loop filtering may be performed.

Meanwhile, when the MVP mode is applied, a reference picture index may be explicitly signaled. In this case, a reference picture index (refidxL0) for L0 prediction and a reference picture index (refidxL1) for L1 prediction may be signaled separately. For example, when the MVP mode is applied and pair prediction is applied, both information about refidxL0 and information about refidxL1 may be signaled.

When the MVP mode is applied, information on the MVD derived from the encoding device 100 may be signaled to the decoding device 200 as described above. The information on the MVD may include, for example, information indicating the absolute value of the MVD and the x and y components of the sign. In this case, whether the absolute MVD value is greater than 0 (abs_mvd_greater0_flag), whether it is greater than 1, and information indicating the remainder of the MVD (abs_mvd_greater1_flag) may be signaled in stages. For example, information indicating whether the absolute MVD value is greater than 1 (abs_mvd_greater1_flag) may be signaled only when the value of the flag information (abs_mvd_greater0_flag) indicating whether the absolute MVD value is greater than 0 is 1.

For example, information on MVD may be configured in a syntax as shown in Table 1 below, encoded in the encoding device 100, and signaled to the decoding device 200.

For example, MVD[compIdx] may be derived based on abs_mvd_greater0_flag[compIdx] *( abs_mvd_minus2[compIdx] + 2) * (1-2 * mvd_sign_flag[compIdx]). Here, compIdx (or cpIdx) represents the index of each component, and may have a value of 0 or 1. compIdx 0 may indicate the x component, and compIdx 1 may indicate the y component. However, this is an example, and values for each component may be expressed using a coordinate system other than the x and y coordinate systems.

Meanwhile, MVD (MVDL0) for L0 prediction and MVD (MVDL1) for L1 prediction may be differentiated and signaled, and the information on MVD may include information on MVDL0 and/or information on MVDL1. For example, when the MVP mode is applied to the current block and bidirectional prediction is applied, both information about MVDLO and information about MVDL1 may be signaled.

Symmetric MVD (SMVD)

Meanwhile, when bidirectional prediction is applied, symmetric MVD (SMVD) may be used in consideration of coding efficiency. In this case, signaling of some of the motion information may be omitted. For example, when SVMD is applied to the current block, information on refidxL0, information on refidxL1, and information on MVDL1 are not signaled from the encoding device 100 to the decoding device 200 and may be derived internally. . For example, when the MVP mode and bidirectional prediction are applied to the current block, flag information indicating whether to apply SMVD (eg, symmetric MVD flag information or sym_mvd_flag syntax element) may be signaled and the value of the flag information is 1 The decoding apparatus 200 may determine that SMVD is applied to the current block.

When the SMVD mode is applied (that is, when the value of symmetric MVD flag information is 1), information about mvp_l0_lfag, mvp_l1_flag, and MVDL0 may be explicitly signaled, and information about refidxL0 as described above , information on refidxL1, and signaling of information on MVDL1 may be omitted, and may be derived inside the decoder. For example, refidxL0 is derived as an index indicating the previous reference picture closest to the current picture in the order of the picture order count (POC) within the reference picture list 0 (which may be referred to as list 0, L0, or the first reference list). Can be. refidxL1 may be derived as an index indicating a subsequent reference picture closest to the current picture in the POC order in reference picture list 1 (which may be referred to as List 1, L1, or a second reference picture list). Also, for example, both refidxL0 and refidxL1 may be derived as 0, respectively. Also, for example, refidxL0 and refidxL1 may be derived as minimum indexes having the same POC difference in relation to the current picture. As a more specific example, [POC of the current picture]-[POC of the first reference picture indicated by refidxL0] is referred to as the first POC difference, and [POC of the second reference picture indicated by refidxL1] is referred to as the second POC difference. In this case, only when the first POC difference and the second POC difference are the same, the value of refidxL0 indicating the first reference picture is derived as the value of refidxL0 of the current block, and the value of refidxL1 indicating the second reference picture is the same as that of the current block. It can also be derived as the value of refidxL1. In addition, for example, when there are a plurality of sets in which the first POC difference and the second POC difference are the same, refidxL0 and refidxL1 of the set with the minimum difference may be derived as refidxL0 and refidxL1 of the current block.

MVDL1 can be derived as -MVDL0. For example, the final MV for the current block may be derived as in Equation 1 below.

In Equation 1, mvx ₀ and mvy ₀ represent the x and y components of the L0 direction motion vector for the current block, and mvx ₁ and mvy ₁ represent the x and y components of the motion vector for L0 direction prediction for the current block. And the x and y components of the motion vector for L1 direction prediction. mvp ₀ and mvp ₀ denote an MVP motion vector for L0 direction prediction (L0 base motion vector), and mvp ₁ and mvp ₁ denote an MVP motion vector for L1 direction prediction (L1 base motion vector). mvd ₀ and mvd ₀ represent the x and y components of MVD for L0 direction prediction. According to Equation 1, MVD for L1 direction prediction has the same value as L0 MVD, but has an opposite sign.

Affine mode

Existing video coding systems used one motion vector to represent the motion of a coding block (using a translation motion model). However, the method using one motion vector may express optimal motion in block units, but it is not actually optimal motion for each pixel, so if the optimal motion vector is determined in pixel units, coding efficiency can be improved. . To this end, the present embodiment describes an affine motion prediction method for encoding/decoding using an affine motion model. In the affine motion prediction method, a motion vector may be expressed in units of each pixel of a block using two, three, or four motion vectors.

The affine motion model can represent four motions as shown in FIG. 16. The affine motion model that expresses three movements (translation, scale, and rotate) among the movements that can be expressed by the affine motion model is referred to as a similar (or simplified) affine motion model. The proposed methods are described based on the affine motion model. However, the embodiments of the present specification are not limited to a similar (or singular) affine motion model.

As shown in FIGS. 21A and 21B, the affine motion prediction may determine a motion vector for each pixel position included in a block using two or more control point motion vectors (CPMVs).

For the 4-parameter affine motion model (FIG. 21A), the motion vector at the sample position (x, y) can be derived as in Equation 2 below.

For a 6-parameter affine motion model (FIG. 21B), a motion vector at the sample position (x, y) can be derived as in Equation 3 below.

Where {v _0x , v _0y } is the CPMV of the CP at the top-left corner of the coding block, and {v _1x , v _1y } is the CPMV of the CP at the top-right corner, {v _2x , v _2y } is the CPMV of the CP at the bottom-left corner. In addition, W corresponds to the width of the current block, H corresponds to the height of the current block, and {v _x , v _y } is a motion vector at the position {x, y}.

In the encoding/decoding process, a motion vector field (MVF), which is an affine, may be determined in a pixel unit or a predefined subblock unit. When the MVF is determined in units of pixels, a motion vector is obtained based on each pixel value, and when the MVP is determined in units of sub-blocks, the center of the sub-block (the lower right of the center, that is, the lower right of the center 4 samples) pixel A motion vector of a corresponding block may be obtained based on a value. In the following description, as shown in FIG. 22, it is assumed that the affine MVF is determined in units of 4*4 subblocks. However, this is only for convenience of explanation, and the size of the subblock may be variously changed.

That is, when affine prediction is available, motion models applicable to the current block may include the following three types. Translational motion model, 4-parameter affine motion model, and 6-parameter affine motion mode. Here, the translational motion model can represent a model in which an existing block-based motion vector is used, a 4-parameter affine motion model can represent a model in which two CPMVs are used, and a 6-parameter affine motion model can represent three It can indicate the model in which CPMV is used.

The affine motion prediction may include an affine MVP (or affine inter) mode and an affine merge. In affine motion prediction, motion vectors of a current block may be derived in units of samples or sub-blocks.

Affine merge

In the afine merge mode, the CPMV may be determined according to the afine motion model of the neighboring block coded by the afine motion prediction. Affine-coded neighboring blocks in the search order may be used for the affine merge mode. When one or more neighboring blocks are coded by affine motion prediction, the current block may be coded as AF_MERGE. That is, when the affine merge mode is applied, CPMVs of the current block may be derived using CPMVs of neighboring blocks. In this case, CPMVs of the neighboring block may be used as CPMVs of the current block, or CPMVs of the neighboring block may be used as CPMVs of the current block by being modified based on the size of the neighboring block and the size of the current block.

When the affine merge mode is applied, an affine merge candidate list may be constructed to derive CPMVs for the current block. The affine merge candidate list may include at least one of the following candidates, for example.

1) Inherited affine candidates

2) Constructed affine candidates

3) zero MV candidate

Here, the inherited affine candidates are candidates derived based on the CPMVs of the neighboring block when the neighboring block is coded in the affine mode, and constructed affine candidates are the corresponding CP neighboring blocks in each CPMV unit. It is a candidate derived by constructing CPMVs based on MV, and a zero MV candidate may represent a candidate composed of CPMVs having a value of 0.

Referring to FIG. 23, a coding device (encoding device or decoding device) inserts inherited affine candidates into a candidate list (S2310), and constructs constructed affine candidates into an affine candidate list. Then, a zero MV candidate may be inserted into the affine candidate list (S2330). In an embodiment, when the number of candidates included in the candidate list is smaller than the reference number (eg, two), the coding apparatus may insert the configured affine candidates or the zero MV candidate.

24 shows examples of blocks for deriving an inherited affine motion predictor according to an embodiment of the present specification, and FIG. 25 is a diagram for deriving an inherited affine motion predictor according to an embodiment of the present specification. An example of control point motion vectors is shown.

There may be up to two (one from the left neighboring CU and one of the upper neighboring CUs) of inherited affine candidates, which may be derived from the affine motion model of neighboring blocks. Candidate blocks are shown in FIG. 24. The scan order for the left predictor is A0-A1, and the scan order for the upper predictor is B0-B1-B2. Only the first inherited candidates from each side are selected. A pruning check may not be performed between the two inherited candidates. When the adjacent affine CU is identified, control point motion vectors of the adjacent affine CU may be used to derive a control point motion vector predictor (CPMVP) candidate from the affine merge list of the current CU. As shown in FIG. 25, if the left neighboring block A is coded in the afine mode, motion vectors of the CU including the block A are v ₂ , v ₃ of the upper left corner, the upper right corner, and the lower left corner. , And v ₄ are used. When block A is coded with a 4-parameter affine model, two CPMVs of the current CU are calculated according to v ₂ and v ₃ . When block A is coded with a 6-parameter model, the three CPMVs of the current CU are calculated according to v ₂ , v ₃ , and v ₄ .

The constructed affine merge means a candidate formed by combining neighboring translational motion information for each control point. As shown in FIG. 26, motion information for control points is derived from specified spatial and temporal neighbors. CPMV _k (k = 1, 2, 3, 4) represents the kth control point. For CPMV1 (CP0) of the upper left corner, blocks are checked in the order of B2-B3-A2 and the MV of the first available block is used. Blocks are checked in the order of B1-B0 with respect to CPMV2 (CP1) in the upper right corner, and blocks in the order of A1-A0 with CPMV3 (CP2) in the lower left corner. If available, TMVP is used for CPMV4 (CP3) in the lower right corner.

When MVs of the four control points are obtained, affine merge candidates are configured based on this motion information. The following combinations of control point MVs are used in order:

{CPMV1, CPMV2, CPMV3}, {CPMV1, CPMV2, CPMV4}, {CPMV1, CPMV3, CPMV4},

{CPMV2, CPMV3, CPMV4}, {CPMV1, CPMV2}, {CPMV1, CPMV3}

Combinations of three CPMVs constitute a 6-parameter affine merge candidate, and a combination of two CPMVs constitutes a 4-parameter affine merge candidate. In order to avoid the motion scaling process, if the reference indices of the control points are different, the combination of the related control point MVs is discarded.

Affine MVP

In the afine MVP mode, after two or more control point motion vector prediction (CPMVP) and CPMV for the current block are determined, a control point motion vector difference (CPMVD) corresponding to the difference value is obtained from the encoding device 100 from the decoding device ( 200).

When the affine MVP mode is applied, an affine MVP candidate list may be configured to derive CPMVs for the current block. For example, the affine MVP candidate list may include at least one of the following candidates. For example, the affine MVP candidate list may include a maximum of n (eg, 2) candidates.

1) Inherited affine mvp candidates that extrapolated from the CPMVs of the neighbor CUs (S2710)

2) Constructed affine mvp candidates CPMVPs that are derived using the translational MVs of the neighbor CUs) derived using the translational MVs of adjacent CUs (S2720)

3) Additional candidates based on Translational MVs from neighboring CUs (S2730)

4) Zero MVs candidate (S2740)

Here, the inherited affine candidate is a candidate derived based on the CPMVs of the neighboring block when the neighboring block is coded in the affine mode, and the constructed affine candidate is each CPMV unit. As a candidate, it is a candidate derived by configuring CPMVs based on the MV of a block adjacent to the corresponding CP, and a zero MV candidate represents a candidate composed of CPMVs whose value is 0. When the maximum number of candidates for the affine MVP candidate list is two, in the above order, 2) or less candidates may be considered and added when the number of current candidates is less than two. In addition, additional candidates based on Translational MVs from neighboring CUs from neighboring CUs may be derived in the following order.

1) If the number of candidates is less than 2 and the constructed candidate CPMV0 is valid, CPMV0 is used as an affine MVP candidate. That is, the MVs of CP0, CP1, and CP2 are all set to be the same as CPMV0 of the constructed candidate.

2) If the number of candidates is less than 2 and the constructed candidate CPMV1 is valid, CPMV1 is used as an affine MVP candidate. That is, the MVs of CP0, CP1, and CP2 are all set to be the same as CPMV1 of the constructed candidate.

3) If the number of candidates is less than 2 and the constructed candidate CPMV2 is valid, CPMV2 is used as the affine MVP candidate. That is, the MVs of CP0, CP1, and CP2 are all set to be the same as CPMV2 of the constructed candidate.

4) If the number of candidates is less than 2, TMVP (temporal motion vector predictor or mvCol) is used as an affine MVP candidate.

The affine MVP candidate list may be derived by the procedure shown in FIG. 27.

The order of checking inherited MVP candidates is the same as that of the inherited affine merge candidates. The difference is that, for the MVP candidate, only affine CUs having the same reference picture as the current block are considered. When the inherited affine motion predictor is added to the candidate list, the pruning process is not applied.

The configured MVP candidate is derived from neighboring blocks shown in FIG. 26. The same confirmation order as the composition of the affine merge candidate is used. In addition, reference picture indexes of neighboring blocks are also checked. The first block that is inter-coded in the check order and has the same reference picture as the current CU is used.

SbTMVP (Subblock-based temporal motion vector prediction)

Subblock-based temporal motion vector prediction (SbTMVP) method may be used. Similar to temporal motion vector prediction (TMVP), SbTMVP may use a motion field in the co-located picture to improve the motion vector predictor and merge mode for CUs in the current picture. The same co-located picture used by TMVP is used for SbTMVP. SbTMVP differs from TMVP in the following two aspects.

1. TMVP predicts motion at the CU level, but SbTMVP predicts motion at the sub-CU level.

2. TMVP The temporal motion vector is obtained from the co-located block in each co-located picture (here, the co-located block is currently the lower right or center block of the CU), whereas SbTMVP retrieves temporal motion information from the co-located picture. Apply motion shift before import, where the motion shift is obtained from a motion vector from one of the spatial neighboring blocks of the current CU.

The SbTMVP process is shown in FIGS. 28A and 28B. SbTMVP predicts motion vectors of sub-CUs in the current CU in two steps. In the first step, the spatial neighbors in FIG. 30A are examined in the order of A1, B1, B0, and A0. When the first spatial neighboring block having a motion vector using the co-located picture as its reference picture is identified, this motion vector is selected to be motion shifted. (As soon as and the first spatial neighboring block that has a motion vector that uses the collocated picture as its reference picture is identified, this motion vector is selected to be the motion shift to be applied). If such motion is not confirmed from spatial surroundings, the motion shift is set to (0, 0).

In the second step, the motion shift identified in the first step is applied to obtain sub-CU level motion information (motion vectors and reference indices) from the co-located picture as shown in FIG. 30B (that is, the current Are added to the block's coordinates). The example in FIG. 30B assumes that a motion shift is set by the motion of block A1. Then, for each sub-CU, the motion information of the corresponding block (the smallest motion grid covering the center sample) in the co-located picture is used to derive the motion information for the sub-CU. After the motion information of the co-located sub-CU is confirmed, it is converted into reference indices and motion vectors of the current sub-CU in a manner similar to the TMVP process, where temporal motion scaling is the reference pictures of the temporal motion vectors. It is applied to be aligned with temporal motion vectors.

The combined subblock-based merge list including the SbTMVP candidate and the afine merge candidate may be used for signaling of the afine merge mode (which may be referred to as a subblock-based merge mode). The SbTMVP mode may be activated/deactivated by a sequence parameter set (SPS). When the SbTMVP mode is activated, the SbTMVP predictor is added as the first entry in the list of subblock-based merge candidates, and affine merge candidates are added later. The maximum allowed size of the affine merge candidate list may be 5.

The size of the sub-CU used in SbTMVP is fixed at 8x8, the same is applied to the affine merge mode, and the SbTMVP mode can only be applied to CUs with both width and height greater than or equal to 8.

The encoding logic of the additional SbTMVP merge candidate is the same as other merge candidates, that is, for each CU in the P or B slice, an additional RD (rate-distortion) check is performed to determine whether to use the SbTMVP candidate.

AMVR (Adaptive Motion Vector Resolution)

Conventionally, when use_integer_mv_flag in a slice header is 0, a motion vector difference (MVD) (between the predicted motion vector and the motion vector of the CU) may be signaled in units of quarter-luma-samples. In this document, the CU-level AMVR scheme is introduced. The AMVR may cause the MVD of the CU to be coded in units of 1/4 luminance samples, integer luminance samples, or 4 luminance samples. If the current CU has at least one non-zero MVD component, a CU-level MVD resolution indicator is conditionally signaled. If all MVD components (i.e., horizontal and vertical MVDs for reference list L0 and reference list L1) are 0, then the 1/4 luminance sample MVD resolution is inferred.

For a CU with at least one non-zero MVD component, a first flag is signaled to determine whether 1/4 luminance sample MVD accuracy is applied for that CU. If the first flag is 0, no additional signaling is required and 1/4 luminance sample MVD accuracy is used for the current CU. Otherwise, a second flag is signaled to indicate whether integer luminance samples or 4 luminance samples MVD accuracy is used. In order for the reconstructed MV to assure the intended accuracy (1/4 luminance sample, integer luminance sample, or 4 luminance sample), the motion vector predictors for CU have the same accuracy as the motion vector predictors previously added with MVD Can be rounded to have Motion vector predictors can be rounded to zero. (I.e., negative motion vector predictors are rounded to positive infinity and positive motion vector predictors are rounded to negative infinity). The encoder determines the motion vector resolution for the current CU using the RD check. In order to avoid always performing three CU-level RD checks for each MVD resolution, the RD check of 4 luminance samples MVD resolution can be called conditionally. The RD cost of 1/4 sample MVD accuracy is calculated first. Then, the RD cost of the integer luminance sample MVD accuracy is compared with the RD cost of the 1/4 luminance sample MVD accuracy to determine whether it is necessary to check the RD cost of the 4 luminance sample MVD accuracy. When the RD cost for 1/4 luminance sample MVD accuracy is less than the RD cost for integer luminance sample MVD accuracy, the RD cost of 4 sample MVD accuracy is omitted.

Motion Field Storage

In order to reduce the memory load, motion information of a reference picture previously decoded may be stored in units of a predetermined area. This may be referred to as temporal motion field storage, motion field compression, or motion data compression. In this case, the storage unit of motion information may be set differently depending on whether the affine mode is applied. In this case, the one with the highest accuracy among explicitly signaled motion vectors is a quarter-luma-sample. In some inter prediction modes, such as the afine mode, motion vectors are derived at 1/16th-luma-sample precision and motion compensated prediction is performed at 1/16th-sample accuracy. In terms of internal motion field storage, all motion vectors are stored with 1/16 luminance sample accuracy.

In this document, for storing the temporal motion field used by TMVP and ATMVP, motion field compression is performed with 8x8 granularity.

History-based merge candidate derivation

HMVP (history-based MVP) merge candidates may be added to the merge list after spatial MVP and TMVP. In this method, motion information of a previously coded block is stored in a table and used as an MVP for a current CU. A table composed of multiple HMVP candidates is maintained during the encoding/decoding process. When a new CTU row is used, the table is reset (emptied). When there is a CU coded by inter prediction other than a subblock, related motion information is added to the last entry of the table as a new HMVP candidate.

In one embodiment, the HMVP table size (S) is set to 6, which means that a maximum of 6 HVMP candidates can be added to the table. When inserting a new motion candidate into a table, a constrained first-in-first-out (FIFO) rule is used. Here, a redundancy check is first performed to check whether an HMVP candidate with the same HMVP candidate to be added exists in the table. If the same HMVP candidate exists, the same existing HMVP candidate is removed from the table and all HMVP candidates are moved in the previous order.

HMVP candidates can be used in the merge candidate list construction process. The most recent HMVP candidates are identified in the table, and are inserted into the merge candidate list in order after TMVP candidates. Redundancy check for HMVP candidates is applied to spatial or temporal merge candidates.

In order to reduce the number of times the redundancy check operations are performed, the following simplification methods may be used.

1) The number of HMVP candidates for creating a merge list is (N <= 4)? M: It is set to (8-N). Here, N denotes the number of candidates existing in the merge list, and M denotes the number of HMVP candidates available in the table.

2) When the total number of usable merge candidates reaches a value minus 1 from the maximum number of allowed merge candidates, the process of constructing a merge candidate list from HVMP is terminated.

Pair-wise average merge candidates derivation

Pair average candidates are generated by an average of predefined pairs of candidates existing in the merge candidate list. Here, the predefined pairs are defined as {(0, 1), (0, 2), (1, 2), (0, 3), (1, 3), (2, 3)}, 0, 1 Numbers such as, 2, and 3 are merge indexes in the merge candidate list. The average of motion vectors is calculated individually for each reference list. If both motion vectors are available in one list, the average value of the two motion vectors is used even if the two motion vectors are for different reference pictures. If only one motion vector is available, the available motion vector is used immediately. If there are no motion vectors available, the list is kept invalid.

When the merge list is not filled even after the pair average merge candidate is added, zero MVs are inserted until the maximum number of merge candidates is reached.

MMVD (Merge mode with MVD)

In addition to a merge mode in which the implicitly derived motion information is directly used to generate a prediction sample of the current CU, a merge mode with motion vector differences may be provided. Since similar motion information duo methods can be used in skip mode and merge mode, MMVD can be applied to skip mode. The MMVD flag may signal the MMVD flag after transmission of the skip flag and merge mode to indicate whether the MMVD mode is used for the CU. In MMVD, after a merge candidate is selected, a motion vector may be refined by the signaled MVD information. The additional information includes a merge candidate flag, an index indicating a motion size, and an index for an indication of a motion direction. In MMVD mode, one of the first two candidates in the merge list is selected as the MV basis. A merge candidate flag is signaled to indicate which candidate is used.

29 illustrates an example of a merge mode with MVD (MMVD) search point according to an embodiment of the present specification. The distance index indicates motion size information and a pre-defined offset from a starting point. As shown in Fig. 29, an offset is added to the horizontal component or the vertical component. The relationship between the distance index and the predefined offset may be defined as shown in Table 2.

The direction index represents the direction of the MVD with respect to the viewpoint. The direction index may represent one of four directions as shown in Table 3. The meaning of the MVD code may vary depending on the information of the starting MV. If the starting MV is a uni-prediction MV or a bi-prediction MV pointing to the same side of the current picture (i.e., when the POCs of both reference pictures are larger than the current picture, or both are smaller than the current picture) ), the code in Table 3 below indicates the code of the MV offset added to the starting MVD. If the starting MV uses two MVs pointing to the other side of the current picture (that is, if the POC of one reference picture is larger than the POC of the current picture and the POC of the other reference picture is smaller than the POC of the current picture), The sign represents the sign of the MV offset added to the list0 MV component of the starting MV, and the sign for list1 MV is the opposite.

DMVR (Decoder side motion vector refinement)

The DMVR is a method of performing prediction by refining motion information of neighboring blocks at the decoder side. When DMVR is applied, the decoder may derive improved motion information through cost comparison based on a template generated by using motion information of neighboring blocks in the skip/merge mode. The DMVR according to the present embodiment can improve the precision of motion prediction and improve compression performance without signaling of additional information. In this embodiment, the description is based on a decoder for convenience of description, but a similar method may be applied to an encoder in the DMVR of this embodiment.

30 shows an example of a DMVR process according to an embodiment of the present specification. Referring to FIG. 30, first, the decoder may generate a template (or a viral template) by weighting (eg, averaging) prediction blocks identified by initial motion vectors (or motion information) in the list 0 and list 1 directions. have. Here, the initial motion vector represents a motion vector derived using motion information of neighboring blocks in the merge/skip mode.

In addition, the decoder may derive a motion vector that minimizes a difference value between the template and the sample region of the reference image through a template matching operation. Here, the sample area represents an area surrounding the initial prediction block in the reference picture, and the sample area may be referred to as a surrounding area, a reference area, a search area, a search range, or a search space. The template matching operation may include calculating a cost measurement value between the template and the sample area of the reference image. For example, the sum of absolute differences (SAD) can be used to measure cost. As an example, a normalized SAD may be used as a cost function. In this case, the matching cost may be given as SAD(T-mean(T), 2 * P[x]-2 * mean(P[x])). Here, T represents a template, and P[x] represents a block in the search area. Further, a motion vector for calculating the minimum template cost for each of the two reference pictures may be considered as an updated motion vector (replaces the initial motion vector). As shown in FIG. 30, the decoder may generate a final bidirectional prediction result (ie, a final bidirectional prediction block) using the updated motion vectors MV0' and MV1'. As an embodiment, a multi-iteration operation for deriving an updated (or new) motion vector may be used to obtain a final bidirectional prediction result.

31 shows an example of a DMVR process according to an embodiment of the present specification. In one embodiment, the decoder may invoke the DMVR process to improve the accuracy of initial motion compensation prediction (ie, motion compensation prediction through a conventional merge/skip mode). For example, when the prediction mode of the current block is the merge mode or the skip mode, and the bidirectional prediction in which the bidirectional reference picture is in the opposite direction based on the current picture in display order is applied to the current block, as shown in FIG. DMVR process can be performed.

Referring to FIG. 31, a decoder may first generate a template (or a viral template) by weighting (eg, averaging) prediction blocks identified by initial motion vectors (or motion information) in the list0 and list1 directions ( S3110). In addition, the decoder may refine the initial motion vector by performing a search in units of integer pixels through template matching (S3120). For example, motion compensation may be performed by using bilinear interpolation filtering in the search step. If the initial motion vector is changed (S3130), the decoder may search in integer pixel units through template matching and refine the changed motion vector (S3140). Even in this case, motion compensation may be performed using bilinear interpolation filtering. Thereafter, the decoder may perform motion compensation using the last motion vector (or the last updated motion vector or the last changed motion vector) (S3150). For example, after the DMVR process is completed, final motion compensation may be performed using a regular interpolation filter. If the initial motion vector is not changed, the decoder may perform motion compensation using the initial motion vector. For example, after the DMVR process is completed, final motion compensation may be performed using regular interpolation filters.

As an embodiment, the search area may be composed of a total of 6 search points including one central point and five surrounding points. In the first step, five points (CENTER, UP, LEFT, DOWN, and RIGHT) may first be set as the search area. In the second step, one of the four vertices (TOP-LEFT, TOP-RIGHT, BOTTOM-LEFT, and BOTTOM-RIGHT) may be included in the search area according to the matching cost acquired for the four neighboring points. For example, in the process of selecting one of four vertices, a gradient of cost may be used. In addition, as an example, the refined motion vector may be updated and stored through a motion vector update process, or may be used only for motion compensation of the current block without a motion vector update process in consideration of hardware design and implementation ease.

32 illustrates an example of a viral matching method according to an embodiment of the present specification. In another embodiment, in performing DMVR, the decoder may use Mean Sum of Absolute Difference (MRSAD) to measure matching cost. In particular, in this embodiment, the decoder may perform motion vector refinement by calculating the MRSAD between prediction samples within two reference pictures without generating a template. 32 shows an embodiment of bilateral matching using MRSAD.

Referring to FIG. 32, the decoder may calculate the matching cost by calculating the MRSAD between blocks identified by motion vectors indicating adjacent pixels of pixels indicated by motion vectors in the list0 and list1 directions. In addition, the decoder may select a search point having the least cost as a pair of processed motion vectors. As an embodiment, after the search region is set, unidirectional prediction may be performed using a regular 8 tap DCTIF interpolation filter. In addition, as an example, 16-bit precision may be used for MRSAD calculation, and clipping and/or rounding operations may not be applied prior to MRSAD calculation in consideration of an internal buffer.

33 illustrates an example of an integer pixel search and a half pixel search according to an embodiment of the present specification. In another embodiment, after completing the integer precision search, the decoder may refine the motion vector by performing a half pixel precision search.

In this embodiment, the integer precision search may be performed by the adaptive pattern method described in FIG. 31 above. First, a cost corresponding to a center point indicated by an initial motion vector may be calculated first. Further, the cost of the other four points (P1, P2, P3, P4) is calculated, and the last sixth point may be selected using a gradient of the previously calculated cost.

By way of example, the output of the DMVR process corresponds to a refined motion vector pair corresponding to the minimum cost. After one iteration, if the minimum cost is achieved at the center point of the search space, the motion vector is not changed and the refinement process can be terminated. If the minimum cost does not correspond to the center point and does not exceed the search area, the best cost point is regarded as the center point and the process can continue. The half sample fine search can be applied only when the search area is not exceeded. in this case. Four MRSAD calculations corresponding to the plus shape points around the center of the point (or pixel) selected as the best (ie, having the best cost) during the integer precision search may be performed. Finally, a pair of refined motion vectors corresponding to the minimum cost point is output.

Generate prediction samples

A predicted block for the current block may be derived based on motion information derived according to the prediction mode. The predicted block may include prediction samples (prediction sample array) of the current block. When the motion vector of the current block indicates a fractional sample unit, an interpolation procedure may be performed. Prediction samples of the current block may be derived from reference samples in units of fractional samples in a reference picture through an interpolation procedure. When afine inter prediction is applied to the current block, prediction samples may be generated based on a motion vector in units of samples/subblocks. When bi prediction is applied, prediction samples derived based on L0 direction prediction (i.e., prediction using a reference picture in an L0 reference picture list and an L0 motion vector) and L1 prediction (i.e., an L1 reference picture list) Prediction samples derived through a weighted sum (according to a phase) or a weighted average of prediction samples derived based on prediction using an internal reference picture and an L1 motion vector) may be used as prediction samples of the current block. When pair prediction is applied, when the reference picture used for L0 prediction and the reference picture used for L1 prediction are located in different temporal directions with respect to the current picture (that is, bi-directional prediction while being pair prediction If applicable) may be referred to as true pair prediction.

As described above, reconstructed samples and reconstructed pictures may be generated based on the derived prediction samples, and then procedures such as in-loop filtering may be performed.

BWA (Bi-prediction with weighted average)

As described above, according to the present specification, when pair prediction is applied to a current block, a prediction sample may be derived based on a weighted average. The pair prediction signal (ie, pair prediction samples) may be derived through a simple average or weighted average of the L0 prediction signal (L0 prediction samples) and the L1 prediction signal (L1 prediction samples). When the prediction sample derivation by a simple average is applied, the pair prediction samples may be derived as average values of the L0 prediction samples based on the L0 reference picture and the L0 motion vector, and the L1 prediction samples based on the L1 reference picture and the L1 motion vector. . According to an embodiment of the present specification, when pair prediction is applied, a pair prediction signal (pair prediction samples) may be derived through a weighted average of the L0 prediction signal and the L1 prediction signal as shown in Equation 4 below.

In Equation 4, P _bi-pred represents a pair prediction sample value, P ₀ represents an L0 prediction sample value, P ₁ represents an L0 prediction sample value, and w represents a weight value.

In weighted average pair prediction, five weight values (w) may be allowed, and the weight values (w) may be -2, 3, 4, 5, 10. For each CU to which pair prediction is applied, the weight w may be determined by one of two methods.

1) For a non-merge CU, the weight index is signaled after MVD.

2) For the merge CU, the weight index is inferred from neighboring blocks based on the merge candidate index.

The weighted sum pair prediction can only be applied to CUs with 256 or more luminance samples (CUs whose product of CU width and CU height is greater than or equal to 256). For low-delay pictures, all five weights can be used. For pictures that are not low-latency, only three weights (3, 4, 5) can be used.

a) In the encoder, a fast search algorithm is applied to find the weight index without significant increase in encoder complexity. These algorithms are summarized below. When combined with AMVR, if the current picture is a low delay picture, only unequal weights are conditionally checked for 1-pel and 4-pel motion vector accuracy.

b) When combined with afine, if the afine mode is currently selected as the optimal mode, affine ME (motion estimation) will be performed for weights that are not the same.

c) When two reference pictures are identical in pair prediction, only weights that are not identical are conditionally checked.

e) According to a POC distance between a current picture and its reference pictures, a coding quantization parameter (QP), and a temporal level, weights that are not identical are not searched when a specific condition is not satisfied.

BDOF (Bi-directional optical flow)

BDOF can be used to process a bi-prediction signal. For example, BDOF can be applied at the 4x4 subblock level. BODF can be applied to CUs that satisfy conditions such as 1), 2), and 3), for example.

1) The height of the CU is not 4, and the CU is not 4x8

2) CU is not coded using afine mode or ATMVP merge mode

3) The CU is coded using the "true" pair prediction mode. That is, one of the reference pictures is located before the current picture in the display order and the other is located after the current picture in the display order.

BDOF can only be applied to the luminance component. Alternatively, BDOF may be applied only to a color difference component, or may be applied to a luminance component and a color difference component. As indicated by its name, the BDOF mode can be based on the optical flow concept, which assumes that the motion of the object is smooth. For 4x4 sub-blocks, motion processing (v _x , v _y ) is calculated by minimizing the difference between the L0 and L1 prediction samples. Motion processing is used to adjust the pair predicted sample values in the 4x4 subblock. The steps below apply to the BDOF process.

First, horizontal and vertical gradients of two prediction signals are calculated by directly calculating the difference between two neighboring samples as shown in Equation 5 below.

Here, I ^(k) (i,j) represents a sample value at the coordinate (i,j) of the prediction signal in the list k (k = 0,1).

The auto- and cross-correlation of the gradients are calculated as in Equations 6 and 7 below.

Where Ω is a 6x6 window surrounding a 4x4 subblock.

Motion processing (v _x , v _y ) is derived as in Equation 8 below using auto- and cross-correlation terms.

In Equation 8,

ego,

Is a floor function.

Based on the motion processing and the gradient, an adjustment value such as Equation 9 below is calculated for each sample in a 4x4 subblock.

Finally, the BDOF samples of the CU are calculated by adjusting the bi-prediction samples as shown in Equation 10 below.

As an example, the values of n _a , n _b , and n _S2 are 3, 6, and 12, respectively. These values are selected so that the multipliers in the BDOF process do not exceed 15 bits and the maximum bit-width of the intermediate parameter in the BDOF process remains within 32 bits.

In order to derive the gradient values, I ^(k) (i,j) needs to be generated in a list of several prediction samples k (k = 0,1) outside of the current CU boundaries.

In FIG. 34, BDOF uses extended rows/columns around CU boundaries. In order to control the computational complexity in the process of generating out-of-boundary prediction samples, a bilinear filter can be used to generate prediction samples in an extended area (area indicated by dotted lines), and an 8-tap motion compensation interpolation filter is used. It is used to generate prediction samples in the CU (area indicated by solid lines). The expanded sample values are used only for gradient calculation. For the remaining steps in the BDOF process, extended sample values are padded from nearest neighbors if any sample and gradient values outside the CU boundaries are needed.

CIIP (combined inter and intra prediction)

CIIP can be applied to the current CU. For example, when a CU is coded in merge mode, if the CU contains at least 64 luminance samples (the product of the CU width and the CU height is greater than or equal to 64), an additional flag indicates whether the CIIP mode is applied to the current CU. May be signaled to indicate. The CIIP mode may also be referred to as a multi-hypothesis mode or an inter/intra multiple hypothesis mode.

Intra prediction mode derivation

Up to four intra prediction modes including DC, PLANAR, HORIZONTAL, and VERTICAL modes can be used to predict the luminance component in the CIIP mode. If the CU shape is very wide (for example, if the width is more than twice the height), the HORIZONTAL mode is not allowed. If the CU shape is very narrow (ie, the height is more than twice the width), the VERTICAL mode is not allowed. For these cases, three intra prediction modes are allowed.

The CIIP mode uses three most probable modes (MPMs) for intra prediction. The CIIP MPM candidate list is formed as follows.

-Left and upper neighboring blocks are set to A and B, respectively

-The prediction modes of block A and block B are named intraModeA and intraModeB, respectively, and are derived as follows.

· Let X be A or B

· If i) Block X is unavailable, ii) Block X is not predicted using CIIP mode, or iii) Block B is located outside the current CTU, intraModeX is set to DC

Otherwise, i) if the intra prediction mode of block X is DC or PLANAR, intraModeX is DC or PLANAR, and ii) the intra prediction mode of block X is "vertical-like" directional mode (greater than 34). Mode), intraModeX is set to VERTICAL, or iii) intraModeX is set to HORIZONTAL if the intra-prediction mode of block X is a "horizontal-like" directional mode (mode less than or equal to 34).

-If intraModeA and intraModeB are the same,

· If intraModeA is PLANAR or DC, 3 MPMs are set in the order of {PLANAR, DC, VERTICAL}

Otherwise, 3 MPMs are set in the order of {intraModeA, PLANAR, DC}

-Otherwise (if intraModeA and intraModeB are not the same),

The first two MPMs are set in the order of {intraModeA, intraModeB}

The uniqueness (redundancy) of PLANAR, DC, VERTICAL is checked for the first two MPM candidates in that order, and if a unique (non-redundant) mode is found, it is added as the third MPM

If the CU shape is very wide or very narrow, the MPM flag is inferred as 1 without signaling. Otherwise, an MPM flag for indicating whether the CIIP intra prediction mode is one of the CIIP MPM candidate modes is signaled.

If the MPM flag is 1, an MPM index indicating which of the MPM candidate modes is used in CIIP intra prediction is additionally signaled. Otherwise, if the MPM flag is 0, the intra prediction mode in the MPM candidate list is set to a "missing" mode. For example, if the PLANAR mode is not in the MPM candidate list, PLANAR becomes the missing mode, and the intra prediction mode is set to PLANAR. Since 4 possible intra prediction modes are allowed in CIIP, the MPM candidate list contains only 3 intra prediction candidates. For color difference components, the DM mode is always applied without additional signaling. That is, the same prediction mode as the luminance component is used for the color difference components. The intra prediction mode of the CU coded with CIIP will be stored and used for intra mode coding of the next neighboring CUs.

Combining the inter and intra prediction signals

The inter prediction signal P _inter in the CIIP mode is derived using the same inter prediction process applied to the general merge mode, and the intra prediction signal P _intra is derived using the CIIP intra prediction according to the intra prediction process. Then, the intra and inter prediction signals are combined using a weighted average, where the weight value depends on the intra prediction mode and where the sample is located in the coding block as follows.

-If the intra prediction mode is DC or planar mode, or the block width or height is less than 4, the same weight is applied to the intra prediction and inter prediction signals.

Otherwise, the weights are determined based on the intra prediction mode (horizontal mode or vertical mode in this case) and the sample position in the block. The horizontal prediction mode will be described as an example (weights for the vertical mode are similar, but can be derived in an orthogonal direction). Set the width of the block to W and the height of the block to H. The coding block is initially divided into 4 co-regional parts, each dimension is (W/4)xH. Starting from the part closest to the intra prediction reference samples and ending the part farthest from the intra prediction samples, the weight wt for each of the four regions is set to 6, 5, 3, and 2. The final CIIP prediction signal may be derived as in Equation 11 below.

In Equation 11, P _CIIP is a CIIP prediction sample value, P _inter is an inter prediction sample value, P _intra is an intra prediction sample value, and wt is a weight.

Example

Hereinafter, a method for adaptively determining whether to perform BDOF in the process of generating an inter prediction sample according to an embodiment of the present specification is provided. For example, embodiments described below provide a method of determining whether to perform BDOF for a block to which an SMVD is applied (SMVD-coded block). First, a method of performing BDOF on an SMVD-coded block and a method of preventing BDOF from being performed on an SMVD-coded block are provided, respectively.

As described above, the BDOF is a process of processing a prediction block in a decoder without separate syntax signaling. The decoder may perform BDOF when a specific condition is satisfied without a separate syntax signaling procedure for determining whether to apply BDOF. In order to determine whether the decoder performs BDOF, a trade-off between the complexity of the decoder stage versus the increase in compression efficiency through improvement of prediction performance may be considered.

Meanwhile, SMVD is a method of performing pair prediction by signaling only L0 and L1 prediction MV candidate indexes (MVP index) and L0 MVD without signaling a reference picture index. Compared to the AMVP method, since coding for the reference picture index and L1 MVD is not performed, SMVD is a method of performing prediction with high accuracy compared to bit reduction. The coding unit syntax structure in which SMVD is reflected according to the present embodiment is shown in Table 4 below.

In Table 4, sym_mvd_flag indicates whether SMVD is applied to the current block (or current CU). If sym_mvd_flag is 1, SMVD is applied to the current block, and if sym_mvd_flag is 0, SMVD is not applied to the current block. In Table 11, conditions for parsing the SMVD flag sym_mvd_flag include a case in which pair prediction is applied to the current block and the affine MVP mode is not applied. In addition, when SMVD is applied, L1 MVD may have the same size as L0 MVD, but may have opposite signs.

Example 1

An embodiment of the present specification provides a method of performing BDOF on an SMVD-coded block in order to improve prediction accuracy while maintaining compression efficiency. As described above, when SMVD is applied, since the reference picture and the L1 MVD are determined without signaling the reference picture index and the L1 MVD, compression efficiency may increase but prediction accuracy may decrease. In this case, prediction accuracy may be improved by applying BDOF to the SMVD-coded block. Below, the decoding procedure for the inter-block according to the embodiment of the present specification is described in the form of a standard document, and the details mean a person skilled in the art. It will be self-explanatory.

This process is called when decoding a coding unit coded in inter prediction mode.

The inputs to this process are as follows.

-A luminance position representing the upper left sample of the current coding block with respect to the upper left luminance sample of the current picture (xCb, yCb)

-Variable cbWidth representing the width of the current coding block in units of luminance samples

-Variable cbHeight representing the height of the current coding block in units of luminance samples

-Variables numSbX and numSbY representing the number of luminance coding subblocks in the horizontal and vertical directions

-xSbIdx = 0 .. numSbX-1 and ySbIdx = 0 .. motion vectors for numSbY-1 mvL0[xSbIdx][ySbIdx] and mvL1[xSbIdx][ySbIdx]

-xSbIdx = 0 .. numSbX-1, and ySbIdx = 0 .. Machined motion vectors for numSbY-1 refMvL0[ xSbIdx ][ ySbIdx] and refMvL1[ xSbIdx ][ ySbIdx]

-Reference indices refIdxL0 and refIdxL1

-xSbIdx = 0 .. numSbX-1 and ySbIdx = 0 .. Prediction list utilization flags for numSbY-1 predFlagL0[xSbIdx][ySbIdx] and predFlagL1[xSbIdx][ySbIdx]

-Pair prediction weight index gbiIdx

-Variable cIdx indicating the color component index of the current block

The output of this process is the array predSamples of prediction samples.

predSamplesL0 _L, predSamplesL1 _L, and placed in arrays (cbWidth) x (cbHeight) of the luminance sample values predicted _{_{_{predSamplesIntra L, predSamplesL0 Cb, predSamplesL1 Cb}}} , predSamplesL0 Cr and predSamplesL1 Cr, predSamplesIntra Cb, and the estimated color difference a predSamplesIntra _Cr Let the arrays of sample values (cbWidth /subWidthC)x(cbHeight / subHeightC).

The width and height subCbWidth and subCbHeight of the current coding subblock as luminance samples are derived as follows.

sbWidth = cbWidth / numSbX

sbHeight = cbHeight / numSbY

For each subblock (xSbIdx, ySbIdx) in the subblock index having xSbIdx = 0 .. numSbX-1, and ySbIdx = 0 .. numSbY-1, the following procedures 1 to 6 are applied.

1. The luminance positions (xSb, ySb) representing the upper left sample of the current coding subblock with respect to the upper left luminance sample of the current picture are derived as follows.

(xSb, ySb) = (xCb + xSbIdx * sbWidth, yCb + ySbIdx * sbHeight)

2. The variable currPic represents the current picture and the variable bdofFlag is derived by the following procedures 1-1 and 1-2.

1-1. When all of the following conditions a to i are satisfied, bdofFlag is set to TRUE.

a. sps_bdof_enabled_flag is 1

b. predFlagL0[ xSbIdx ][ ySbIdx] and predFlagL1[ xSbIdx ][ ySbIdx] are both 1

c. DiffPicOrderCnt( currPic, RefPicList[ 0 ][ refIdxL0]) * DiffPicOrderCnt( currPic, RefPicList[ 1 ][ refIdxL1]) is less than 0

d. MotionModelIdc[ xCb ][ yCb] is 0

e. merge_subblock_flag[ xCb ][ yCb] is 0

f. GbiIdx[ xCb ][ yCb] is 0

g. cIdx is 0

h. cbHeight is greater than or equal to 8

i. cbHeight*cbWidth is greater than or equal to 64

1-2. Otherwise, bdofFlag is set to FALSE.

For X being each of 0 and 1, when predFlagLX[ xSbIdx ][ ySbIdx] is equal to 1, the following applies:

3. For X corresponding to 0 and 1 respectively, when predFlagLX[xSbIdx][ySbIdx] is 1, the following procedures 3-1 to 3-6 are applied.

3-1. Using X and refIdxLX as inputs, a reference picture refPicLX _L composed of an ordered two-dimensional array of luminance samples and ordered two-dimensional arrays refPicLX _Cb and refPicLX _Cr of chrominance samples are derived.

3-2. The motion vector offset mvOffset is set to refMvLX[xSbIdx][xSbIdx]-mvLX[xSbIdx][ySbIdx].

3-3. If one or more of the following conditions is'true', mvOffset[0] is set to 0.

-xSb is not xCb and mvOffset[ 0] is less than 0

-(xSb + sbWidth) is not (xCb + cbWidth) and mvOffset[ 0] is greater than 0

3-4. If one or more of the following conditions is'true', mvOffset[1] is set to 0.

-ySb is not yCb and mvOffset[ 1] is less than 0

-(ySb + sbHeight) is not (yCb + cbHeight) and mvOffset[ 1] is greater than 0

3-4. If cIdx is 0, the following process is applied.

-Luminance position (xCb, yCb), coding subblock width sbWidth in luminance samples, coding subblock height sbHeight in luminance samples, luminance motion vector offset mvOffset, processed luminance motion vector refMvLX[xSb][xSb], reference array The array predSamplesLX _L is derived by calling the fractional sample interpolation process with refPicLX _L , bdofFlag, and cIdx as inputs.

3-5. If cIdx is 1, the following process is applied.

-Luminance position (xCb, yCb), coding subblock width sbWidth / subWidthC, coding subblock height sbHeight / subHeightC, color difference motion vector offset mvOffset, processed color difference motion vector refMvLX[ xSb ][ xSb ], reference array refPicLX _Cb , bdofFlag The array predSamplesLX _Cb is derived by calling the fractional sample interpolation process with, and cIdx as inputs.

3-6. If cIdx is 2, the following process applies.

-Luminance position (xCb, yCb ), coding subblock width sbWidth / subWidthC, coding subblock height sbHeight / subHeightC, color difference motion vector offset mvOffset, processed color difference motion vector refMvLX[ xSb ][ xSb ], reference array refPicLXCr, bdofFlag, And cIdx as inputs, an array predSamplesLX _Cr is derived by calling a fractional sample interpolation process.

4. If bdofFlag is TRUE, procedures 4-1 to 4-4 apply.

4-1. The variable shift is set to Max( 2, 14-BitDepthY ).

4-2. cuDiffThres, bdofBlkDiffThres, and cuSumDiff are derived as in Equation 12 below.

4-3. For xIdx = 0..(sbWidth >> 2)-1 and yIdx = 0..( sbHeight >> 2)-1, the variables bdofBlkSumDiff and the BDOF utilization flag bdofUtilizationFlag[xIdx][yIdx] are the following Equation 13 Is derived as

4-4. If cuSumDiff is less than cuDiffThres, bdofFlag is set to FALSE.

5. The array predSamples of predicted samples are derived by the following procedures 5-1 to 5-3.

5-1. If cIdx is 0, prediction samples predSamples[ xL + xSb ][ yL + ySb] (xL = 0..sbWidth-1, yL = 0..sbHeight-1) in the current luminance coding subblock are derived as follows.

-If bdofFlag is TRUE, nCbW set to luminance coding subblock width sbWidth, nCbH set to luminance coding subblock height sbHeight, predSamplesL0 _L and predSamplesL1 _L , xIdx = 0..( sbWidth >> 2)-1, yIdx = 0. .( sbHeight >> 2)-Variables for 1 predFlagL0[ xSbIdx ][ ySbIdx ], predFlagL1[ xSbIdx ][ ySbIdx ], refIdxL0, refIdxL1, bdofUtilizationFlag[ xIdx ][ yIdx] as input, x _L + The BDOF sample prediction process that outputs xSb ][ y _L + ySb] is called.

-Otherwise (if bdofFlag is FALSE), luminance coding block width sbWidth, luminance coding block height sbHeight and sample arrays predSamplesL0 _L and predSamplesL1 _L , and variables predFlagL0[ xSbIdx ][ ySbIdx ], predFlagL1[ xSbIdx ], refIdxL0, by the refIdxL1, gbiIdx, and cIdx as input, the weighted sample prediction process for outputting _{predSamples [x L + xSb] [} y L + ySb] is called.

5-2. Otherwise, if cIdx is 1, prediction samples in the current coding block component Cb coding block predSamples [xC + xCb / subWidthC ][ yC + yCb / subHeightC] (xC = 0..cbWidth / subWidthC-1, yC = 0. .cbHeight / subHeightC-1) is derived by a weighted sample prediction process, where coding block width nCbW set to cbWidth / subWidthC, coding block height nCbH set to cbHeight / subHeightC, sample arrays predSamplesL0 _Cb and predSamplesL1 _Cb , and variables S predFlagL0[xSbIdx][ySbIdx], predFlagL1[xSbIdx][ySbIdx], refIdxL0, refIdxL1, gbiIdx, and cIdx are input.

5-3. Otherwise, if cIdx is 2, prediction samples in the current coding block component Cb coding block predSamples [xC + xCb / subWidthC ][ yC + yCb / subHeightC] (xC = 0..cbWidth / subWidthC-1, yC = 0. .cbHeight / subHeightC-1) is derived by the weighted sample prediction process, where coding block width nCbW set to cbWidth / subWidthC, coding block height nCbH set to cbHeight / subHeightC, sample arrays predSamplesL0 _Cr and predSamplesL1 _Cr , and variables S predFlagL0[xSbIdx][ySbIdx], predFlagL1[xSbIdx][ySbIdx], refIdxL0, refIdxL1, gbiIdx, and cIdx are input.

6. If cIdx is 0, the following assignments are performed for x = 0..sbWidth-1 and y = 0..sbHeight-1.

MvL0[ xSb + x ][ ySb + y] = mvL0[ xSbIdx ][ ySbIdx]

MvL1[ xSb + x ][ ySb + y] = mvL1[ xSbIdx ][ ySbIdx]

MvDmvrL0[ xSb + x ][ ySb + y] = refMvL0[ xSbIdx ][ ySbIdx]

MvDmvrL1[ xSb + x ][ ySb + y] = refMvL1[ xSbIdx ][ ySbIdx]

RefIdxL0[ xSb + x ][ ySb + y] = refIdxL0

RefIdxL1[ xSb + x ][ ySb + y] = refIdxL1

PredFlagL0[ xSb + x ][ ySb + y] = predFlagL0[ xSbIdx ][ ySbIdx]

PredFlagL1[ xSb + x ][ ySb + y] = predFlagL1[ xSbIdx ][ ySbIdx]

GbiIdx[ xSb + x ][ ySb + y] = gbiIdx

If ciip_flag[xCb][yCb] is 1, the array of prediction samples ciip_flag[xCb][yCb] is modified as follows.

-If cIdx is 0, the following procedures 1) and 2) are applied.

1) A general intra-sample prediction process is called, where the position (xTbCmp, yTbCmp) set to (xCb, yCb), the intra prediction mode predModeIntra set to IntraPredModeY[ xCb ][ yCb], the transform block width nTbW set to cbWidth and cbHeight, and Coding block width nCbW and height nCbH set to height nTbH, cbWidth and cbHeight, variable cIdx are inputs, and (cbWidth)x(cbHeight) array predSamplesIntra _L is output.

2) A weighted sample prediction process for combined merge and intra prediction is called, where sample arrays predSamplesInter and predSamplesIntra, IntraPredModeY[xCb][ yCb] set respectively with coding block width cbWidth, coding block height cbHeight, predSamples and predSamplesIntra _L The intra prediction mode predModeIntra set to] and the color component index cIdx are inputs, and the (cbWidth)x(cbHeight) array predSamples is output.

-Otherwise, if cIdx is 1, the following procedures 1) and 2) are applied.

1) The general intra-sample prediction process is called, and the position set to (xCb/subWidthC, yCb/subHeightC) (xTbCmp, yTbCmp), IntraPredModeY[xCb][yCb] Intra prediction mode predModeIntra, cbWidth/subWidthC and cbHeight/subHeightC The transform block width nTbW and height nTbH set to, cbWidth / subWidthC and coding block width nCbW and height nCbH set to cbHeight / subHeightC are inputs, and the (cbWidth / subWidthC)x(cbHeight / subHeightC) array predSamplesIntra _Cb is output.

2) A weighted sample prediction process for the combined merge and intra prediction is called, where the sample arrays predSamplesInter and predSamplesIntra, IntraPredModeY, respectively set to coding block width cbWidth / subWidthC, coding block height cbHeight / subHeightC, predSamples _Cb and predSamplesIntra _Cb , respectively. The intra prediction mode predModeIntra set to [xCb][yCb] and the color component index cIdx are input, and the (cbWidth / subWidthC)x(cbHeight / subHeightC) array predSamples is output.

-Otherwise, if cIdx is 2, the following procedures 1) and 2) are applied.

1) The general intra-sample prediction process is called, and the position set to (xCb/subWidthC, yCb/subHeightC) (xTbCmp, yTbCmp), IntraPredModeY[xCb][yCb], the intra prediction mode predModeIntra, cbWidth/subWidthC and cbHeight/subHeightC The transform block width nTbW and height nTbH set to, cbWidth / subWidthC and coding block width nCbW and height nCbH set to cbHeight / subHeightC are inputs, and the (cbWidth / subWidthC)x(cbHeight / subHeightC) array predSamplesIntra _Cr is output.

2) The weighted sample prediction process for the combined merge and intra prediction is called, where the sample arrays predSamplesInter and predSamplesIntra, IntraPredModeY, respectively set to coding block width cbWidth / subWidthC, coding block height cbHeight / subHeightC, predSamples _Cr and predSamplesIntra _Cr The intra prediction mode predModeIntra set to [xCb][yCb], the color component index cIdx, is input, and the (cbWidth / subWidthC)x(cbHeight / subHeightC) array predSamples is output.

Example 2

This embodiment provides a method of restricting BDOF from being performed on an SMVD-coded block in order to increase compression efficiency versus complexity. As described above, in the SMVD, since the reference picture and the L1 MVD are determined without signaling the reference picture index and the L1 MVD, compression efficiency increases. It was confirmed that when BDOF is performed on the SMVD-coded block, the complexity compared to the compression efficiency may increase. This is because the selection of the SMVD coding method in the encoder itself does not reduce the prediction performance compared to the prediction block of AMVP, and the effect of reducing the number of bits that can be reduced by not performing the reference index and L1 MVD coding is greater. Even if BDOF is applied to such a block, since the increase in prediction performance versus coding complexity is not significant, it is effective not to apply BDOF to a block coded with SMVD. For example, when the SMVD flag (sym_mvd_flag) for the current block is 1, BDOF is not applied to the current block, and when the SMVD flag is 0, BDOF can be applied to the current block. Below, the decoding procedure for the inter-block according to the embodiment of the present specification is described in the form of a standard document, and the details mean a person skilled in the art. It will be self-explanatory.

The inputs to this process are as follows.

-Reference indices refIdxL0 and refIdxL1

-Pair prediction weight index gbiIdx

-Variable cIdx indicating the color component index of the current block

The output of this process is the array predSamples of prediction samples.

sbWidth = cbWidth / numSbX

sbHeight = cbHeight / numSbY

(xSb, ySb) = (xCb + xSbIdx * sbWidth, yCb + ySbIdx * sbHeight)

1-1. When all of the conditions a to j below are satisfied, bdofFlag is set to'TRUE'.

a. sps_bdof_enabled_flag is 1

b. predFlagL0[ xSbIdx ][ ySbIdx] and predFlagL1[ xSbIdx ][ ySbIdx] are both 1

d. MotionModelIdc[ xCb ][ yCb] is 0

e. merge_subblock_flag[ xCb ][ yCb] is 0

f. GbiIdx[ xCb ][ yCb] is 0

g. cIdx is 0

h. sym_mvd_flag[xCb][yCb]가 1이 아님 (sym_mvd_flag[xCb][yCb]가 0임)h. sym_mvd_flag[xCb][yCb] is not 1 (sym_mvd_flag[xCb][yCb] is 0)

i. cbHeight is greater than or equal to 8

j. cbHeight*cbWidth is greater than or equal to 64

1-2. Otherwise, bdofFlag is set to FALSE.

-xSb is not xCb and mvOffset[ 0] is less than 0

-(xSb + sbWidth) is not (xCb + cbWidth) and mvOffset[ 0] is greater than 0

-ySb is not yCb and mvOffset[ 1] is less than 0

-(ySb + sbHeight) is not (yCb + cbHeight) and mvOffset[ 1] is greater than 0

3-4. If cIdx is 0, the following process is applied.

3-5. If cIdx is 1, the following process is applied.

3-6. If cIdx is 2, the following process applies.

4. If bdofFlag is TRUE, procedures 4-1 to 4-4 apply.

4-1. The variable shift is set to Max( 2, 14-BitDepthY ).

4-2. cuDiffThres, bdofBlkDiffThres, and cuSumDiff are derived as in Equation 14 below.

4-3. For xIdx = 0..(sbWidth >> 2)-1 and yIdx = 0..( sbHeight >> 2)-1, the variables bdofBlkSumDiff and the BDOF utilization flag bdofUtilizationFlag[xIdx][yIdx] are the following Equation 15 Is derived as.

4-4. If cuSumDiff is less than cuDiffThres, bdofFlag is set to FALSE.

5-2. Otherwise, if cIdx is 1, prediction samples in the current coding block component Cb coding block predSamples [xC + xCb / subWidthC ][ yC + yCb / subHeightC] (xC = 0..cbWidth / subWidthC-1, yC = 0. .cbHeight / subHeightC-1) is derived by the weighted sample prediction process, where coding block width nCbW set to cbWidth / subWidthC, coding block height nCbH set to cbHeight / subHeightC, sample arrays predSamplesL0 _Cb and predSamplesL1 _Cb , and variables S predFlagL0[xSbIdx][ySbIdx], predFlagL1[xSbIdx][ySbIdx], refIdxL0, refIdxL1, gbiIdx, and cIdx are input.

MvL0[ xSb + x ][ ySb + y] = mvL0[ xSbIdx ][ ySbIdx]

MvL1[ xSb + x ][ ySb + y] = mvL1[ xSbIdx ][ ySbIdx]

MvDmvrL0[ xSb + x ][ ySb + y] = refMvL0[ xSbIdx ][ ySbIdx]

MvDmvrL1[ xSb + x ][ ySb + y] = refMvL1[ xSbIdx ][ ySbIdx]

RefIdxL0[ xSb + x ][ ySb + y] = refIdxL0

RefIdxL1[ xSb + x ][ ySb + y] = refIdxL1

PredFlagL0[ xSb + x ][ ySb + y] = predFlagL0[ xSbIdx ][ ySbIdx]

PredFlagL1[ xSb + x ][ ySb + y] = predFlagL1[ xSbIdx ][ ySbIdx]

GbiIdx[ xSb + x ][ ySb + y] = gbiIdx

-If cIdx is 0, the following procedures 1) and 2) are applied.

-Otherwise, if cIdx is 1, the following procedures 1) and 2) are applied.

2) The weighted sample prediction process for the combined merge and intra prediction is called, where the sample arrays predSamplesInter and predSamplesIntra, IntraPredModeY, respectively set to coding block width cbWidth / subWidthC, coding block height cbHeight / subHeightC, predSamples _Cb and predSamplesIntra _Cb , respectively. The intra prediction mode predModeIntra set to [xCb][yCb], the color component index cIdx, is input, and the (cbWidth / subWidthC)x(cbHeight / subHeightC) array predSamples is output.

-Otherwise, if cIdx is 2, the following procedures 1) and 2) are applied.

1) The general intra-sample prediction process is called, and the position set to (xCb/subWidthC, yCb/subHeightC) (xTbCmp, yTbCmp), IntraPredModeY[xCb][yCb] Intra prediction mode predModeIntra, cbWidth/subWidthC and cbHeight/subHeightC The transform block width nTbW and height nTbH set to, cbWidth / subWidthC and coding block width nCbW and height nCbH set to cbHeight / subHeightC are inputs, and the (cbWidth / subWidthC)x(cbHeight / subHeightC) array predSamplesIntra _Cr is output.

In this embodiment, the condition for performing BDOF includes that the SMVD flag (sym_mvd_flag[xCb][yCb]) of the current block is not 1 (SMVD is not performed). By conditionally performing BDOF in consideration of whether or not SMVD is applied, it is possible to prevent an additional increase in coding complexity for an SMVD-coded block.

Bitstream

The encoded information (eg, encoded video/video information) derived by the encoding apparatus 100 based on the above-described embodiments of the present specification may be output in a bitstream form. The encoded information may be transmitted or stored in a bitstream form in units of network abstraction layer (NAL) units. The bitstream may be transmitted over a network or may be stored in a non-transitory digital storage medium. In addition, as described above, the bitstream is not directly transmitted from the encoding device 100 to the decoding device 200, but may be provided with a streaming/download service through an external server (eg, a content streaming server). Here, the network may include a broadcasting network and/or a communication network, and the digital storage medium may include various storage media such as USB, SD, CD, DVD, Blu-ray, HDD, and SSD.

35 illustrates an example of a video signal encoding procedure for inter prediction according to an embodiment of the present specification. The operations of FIG. 35 may be performed by the inter prediction unit 180 of the encoding apparatus 100 or the processor 510 of the video signal processing apparatus 500.

In step S3510, the encoder generates prediction information for the current block. For example, the encoder may determine an optimal prediction type/mode for the current block in consideration of the RD cost among various prediction types/modes for the current block. For example, the encoder may determine whether to apply SMVD to a current block to which pair prediction is applied, and encode an SMVD flag (sym_mvd_flag) indicating whether the SMVD is applied as prediction information.

When SMVD is applied, a second MVD (L1 MVD) for prediction in the second direction (L1) is determined based on the first MVD (L0 MVD) for prediction in the first direction (L0). More specifically, when SMVD is applied, the size of the second MVD may be the same as the size of the first MVD, and the code of the second MVD may be opposite to that of the first MVD. In addition, when SMVD is applied, information for coding the second MVD and a reference picture index are not signaled. The decoder can determine the second MVD and the reference picture without signaling the reference picture index and information for coding the second MVD. The second MVD is determined based on the first MVD, and the first reference picture (L0 reference picture) for first direction prediction is displayed in the first reference picture list (L0 reference picture list) for the first direction prediction. It is determined as the previous reference picture closest to the current picture, and the second reference picture (L1 reference picture) for second direction prediction is the next closest to the current picture in the display order in the second reference picture list for second direction prediction. It can be determined as a picture.

In step S3520, the encoder generates a prediction sample of the current block based on the prediction information. Here, the encoder checks a condition for the BDOF based on the SMVD flag, and generates a prediction sample of the current block based on the condition for the BDOF.

In an embodiment, when the SMVD flag is 0 and other conditions are satisfied, the condition for the BDOF is determined to be satisfied, and when the SMVD flag is 1, it may be determined that the condition for the BDOF is not satisfied. For example, as described in Example 2, the condition for determining whether to perform BDOF is that the SMVD flag (sym_mvd_flag[xCb][yCb]) is 0 (the current block must be a block not coded with SMVD). ), BDOF can be applied when the SMVD flag sym_mvd_flag[xCb][yCb] is 0 and all other conditions (conditions a to j of procedure 1-1 of the second embodiment) are satisfied. If the SMVD flag (sym_mvd_flag[xCb][yCb]) is 1, BDOF is not applied.

In an embodiment, when the condition for BDOF is satisfied, a prediction sample of a current block is a first prediction sample according to the first direction prediction, a second prediction sample according to the second direction prediction, and a BDOF offset in the current block. Is determined based on, wherein the BDOF offset is to be derived based on a first horizontal gradient and a first vertical gradient for the first prediction sample, and a second horizontal gradient and a second vertical gradient for the second prediction sample. I can.

In an embodiment, if the condition for BDOF is not satisfied, a prediction sample is a first prediction sample (L0 prediction sample) according to the first direction prediction and a second prediction sample (L1 prediction) according to the second direction prediction in the current block. Sample) can be determined based on the weighted sum.

36 illustrates an example of a video signal decoding procedure for inter prediction according to an embodiment of the present specification. The operations of FIG. 36 may be performed by the inter prediction unit 260 of the decoding apparatus 100 or the processor 510 of the video signal processing apparatus 500.

In step S3610, the decoder acquires the SMVD flag of the current block (S3610). The SMVD flag is a flag indicating whether SMVD is applied to the current block.When SMVD is applied, the second direction (L1) is predicted based on the first MVD (L0 MVD) for the first direction (L0) prediction. The second MVD (L1 MVD) is determined. More specifically, when SMVD is applied, the size of the second MVD may be the same as the size of the first MVD, and the code of the second MVD may be opposite to that of the first MVD. In addition, when SMVD is applied, information for coding the second MVD and a reference picture index are not signaled. The decoder can determine the second MVD and the reference picture without signaling the reference picture index and information for coding the second MVD. The second MVD is determined based on the first MVD, and the first reference picture for the first direction prediction (L0 reference picture) is displayed in the first reference picture list (L0 reference picture list) for the first direction prediction. It is determined as the previous reference picture closest to the current picture, and the second reference picture (L1 reference picture) for second direction prediction is the next closest to the current picture in the display order in the second reference picture list for second direction prediction. It can be determined as a picture.

In step S3620, the decoder checks the condition for the BDOF based on the SMVD flag. In an embodiment, when the SMVD flag is 0 and other conditions are satisfied, the condition for the BDOF is determined to be satisfied, and when the SMVD flag is 1, it may be determined that the condition for the BDOF is not satisfied. For example, as described in Embodiment 2, the condition for determining whether to perform BDOF is that the SMVD flag (sym_mvd_flag[xCb][yCb]) should be 0 (the current block should be a block not coded with SMVD). ), BDOF may be applied when the SMVD flag sym_mvd_flag[xCb][yCb] is 0 and all other conditions (conditions a to j of procedure 1-1 of the second embodiment) are satisfied. If the SMVD flag (sym_mvd_flag[xCb][yCb]) is 1, BDOF is not applied.

In step S3630, the decoder generates a prediction sample of the current block based on the condition for the BDOF.

As described above, the embodiments described herein may be implemented and performed on a processor, microprocessor, controller, or chip. For example, the functional units illustrated in each drawing may be implemented and executed on a computer, processor, microprocessor, controller, or chip.

The video signal processing apparatus 500 according to the embodiment of the present specification may include a memory 520 for storing a video signal, and a processor 510 coupled to the memory 520.

In order to encode a video signal, the processor 510 is configured to generate prediction information for a current block and to generate a prediction sample of the current block based on the prediction information. Here, the prediction information includes an SMVD flag related to whether a second MVD for second direction prediction is determined based on a first motion vector difference (MVD) for first direction prediction. In order to generate a prediction sample of the current block, the processor 510 checks a condition for a bi-directional optical flow (BDOF) based on the SMVD flag, and a condition of the current block based on the condition for the BDOF. It is set to generate predictive samples.

In an embodiment, if the condition for the BDOF is not satisfied, the prediction sample is a weighted sum of a first prediction sample according to the first direction prediction and a second prediction sample according to the second direction prediction in the current block. Can be determined based on

In an embodiment, if the SMVD flag is 1, the first reference picture for the first direction prediction is determined as a previous reference picture closest to the current picture in display order in the first reference picture list for the first direction prediction. The second reference picture for the second direction prediction may be determined as a next picture closest to the current picture in the display order in the second reference picture list for the second direction prediction.

For decoding a video signal, the processor 510 obtains an SMVD flag of a current block, wherein the SMVD flag is used for second direction prediction based on a first motion vector difference (MVD) for first direction prediction. It is related to whether the second MVD is determined, and is configured to check a condition for a bi-directional optical flow (BDOF) based on the SMVD flag, and to generate a predicted sample of the current block based on the condition for the BDOF. do.

Further, the processing method to which the present specification is applied may be produced in the form of a program executed by a computer, and may be stored in a computer-readable recording medium. Multimedia data having a data structure according to the present specification may also be stored in a computer-readable recording medium. The computer-readable recording medium includes all kinds of storage devices and distributed storage devices in which computer-readable data is stored. The computer-readable recording medium includes, for example, Blu-ray disk (BD), universal serial bus (USB), ROM, PROM, EPROM, EEPROM, RAM, CD-ROM, magnetic tape, floppy disk, and optical It may include a data storage device. Further, the computer-readable recording medium includes media implemented in the form of a carrier wave (for example, transmission through the Internet). In addition, the bitstream generated by the encoding method may be stored in a computer-readable recording medium or transmitted through a wired or wireless communication network.

In addition, an embodiment of the present specification may be implemented as a computer program product using a program code, and the program code may be executed in a computer according to the embodiment of the present specification. The program code may be stored on a carrier readable by a computer.

A non-transitory computer-readable medium according to an embodiment of the present specification stores one or more instructions executed by one or more processors. The one or more instructions generate prediction information for a current block in order to encode a video signal, and generate a prediction sample of the current block based on the prediction information. The prediction information includes an SMVD flag related to whether a second MVD for second direction prediction is determined based on a first motion vector difference (MVD) for first direction prediction. The one or more instructions determine a condition for a bi-directional optical flow (BDOF) based on the SMVD flag to generate a prediction sample of the current block, and the current condition based on the condition for the BDOF. The video signal processing apparatus 500 (or the encoding apparatus 100) is controlled to generate a prediction sample of a block.

In addition, the one or more instructions obtain an SMVD flag of a current block for decoding a video signal, wherein the SMVD flag is a second motion vector difference (MVD) for first direction prediction. It is related to whether or not a second MVD for direction prediction is determined, a condition for a bi-directional optical flow (BDOF) is checked based on the SMVD flag, and a prediction sample of the current block based on the condition for the BDOF The video signal processing apparatus 500 (or the decoding apparatus 200) is controlled to generate a.

In addition, the decoder and encoder to which the present specification is applied include a multimedia broadcasting transmission/reception device, a mobile communication terminal, a home cinema video device, a digital cinema video device, a surveillance camera, a video chat device, a real-time communication device such as video communication, a mobile streaming device, Storage media, camcorders, video-on-demand (VoD) service providers, OTT video (Over the top video) devices, Internet streaming service providers, three-dimensional (3D) video devices, video telephony video devices, and medical video devices. And can be used to process video signals or data signals. For example, an OTT video (Over the top video) device may include a game console, a Blu-ray player, an Internet-connected TV, a home theater system, a smartphone, a tablet PC, and a digital video recorder (DVR).

The embodiments described above are those in which components and features of the present specification are combined in a predetermined form. Each component or feature should be considered optional unless explicitly stated otherwise. Each component or feature may be implemented in a form that is not combined with other components or features. In addition, it is possible to configure the embodiments of the present specification by combining some components and/or features. The order of operations described in the embodiments of the present specification may be changed. Some configurations or features of one embodiment may be included in other embodiments, or may be replaced with corresponding configurations or features of other embodiments. It is obvious that the embodiments may be configured by combining claims that do not have an explicit citation relationship in the claims or may be included as new claims by amendment after filing.

The embodiments according to the present specification may be implemented by various means, for example, hardware, firmware, software, or a combination thereof. In the case of implementation by hardware, an embodiment of the present specification includes one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), and FPGAs ( field programmable gate arrays), processors, controllers, microcontrollers, microprocessors, etc.

In the case of implementation by firmware or software, an embodiment of the present specification may be implemented in the form of a module, procedure, function, etc. that perform the functions or operations described above. The software code can be stored in a memory and driven by a processor. The memory may be located inside or outside the processor, and may exchange data with the processor through various known means.

In this document, "/" and "," are interpreted as "and/or". For example, "A/B" is interpreted as "A and/or B", and "A, B" is interpreted as "A and/or B". Additionally, “A/B/C” means “at least one of A, B and/or C”. In addition, "A, B, C" also means "at least one of A, B and/or C". (In this document, the term "/" and "," should be interpreted to indicate "and/or." For instance, the expression "A/B" may mean "A and/or B." Further, "A, B" may mean "A and/or B." Further, "A/B/C" may mean "at least one of A, B, and/or C." Also, "A/B/C" may mean " at least one of A, B, and/or C.")

Additionally, in this document "or" is interpreted as "and/or". For example, "A or B" may mean only 1) "A", 2) only "B", or 3) "A and B". In other words, "or" in this document may mean "additionally or alternatively". (Further, in the document, the term "or" should be interpreted to indicate "and/or." For instance, the expression "A or B" may comprise 1) only A, 2) only B, and/or 3) both A and B. In other words, the term "or" in this document should be interpreted to indicate "additionally or alternatively.")

It is apparent to those skilled in the art that the embodiments of the present specification may be embodied in other specific forms without departing from essential features of the present specification. Therefore, the above detailed description should not be construed as restrictive in all respects, but should be considered as illustrative. The scope of this specification should be determined by reasonable interpretation of the appended claims, and all changes within the equivalent scope of this specification are included in the scope of this specification.

Above, the preferred embodiments of the present specification described above are disclosed for purposes of illustration, and those skilled in the art improve and change various other embodiments within the technical spirit and scope of the present specification disclosed in the appended claims below. , Substitution or addition may be possible.

Claims

In the decoding method of a video signal for inter prediction,

Obtaining a symmetric motion vector difference (SMVD) flag of the current block, wherein the SMVD flag is determined by a second MVD for second direction prediction based on a first motion vector difference (MVD) for first direction prediction Is related to whether or not,

Checking a condition for a bi-directional optical flow (BDOF) based on the SMVD flag, and

And generating a prediction sample of the current block based on the condition for the BDOF.
The method of claim 1,

If the SMVD flag is 0 and other conditions are satisfied, the condition for the BDOF is determined to be satisfied,

When the SMVD flag is 1, it is determined that the condition for the BDOF is not satisfied.
The method of claim 2,

When the condition for the BDOF is satisfied, the prediction sample of the current block is based on a first prediction sample according to the first direction prediction in the current block, a second prediction sample according to the second direction prediction, and a BDOF offset. Is determined,

The BDOF offset is derived based on a first horizontal gradient and a first vertical gradient for the first prediction sample, and a second horizontal gradient and a second vertical gradient for the second prediction sample. .
The method of claim 2,

If the condition for the BDOF is not satisfied, the prediction sample is determined based on a weighted sum of a first prediction sample according to the first direction prediction and a second prediction sample according to the second direction prediction in the current block. Decoding method characterized by.
The method of claim 1,

If the SMVD flag is 1,

The size of the second MVD is the same as the size of the first MVD,

The code of the second MVD is opposite to that of the first MVD.
The method of claim 1,

If the SMVD flag is 1,

The first reference picture for first direction prediction is determined as a previous reference picture closest to the current picture in display order in the first reference picture list for first direction prediction,

And the second reference picture for second direction prediction is determined as a next picture closest to the current picture in the display order in the second reference picture list for the second direction prediction.
In the encoding method of a video signal for inter prediction,

Generating prediction information for the current block; and

Generating a prediction sample of the current block based on the prediction information,

The prediction information includes a symmetric motion vector difference (SMVD) flag related to whether a second MVD for second direction prediction is determined based on a first motion vector difference (MVD) for first direction prediction,

Generating a prediction sample of the current block,

Checking a condition for a bi-directional optical flow (BDOF) based on the SMVD flag, and

And generating a prediction sample of the current block based on the condition for the BDOF.
The method of claim 7,

If the SMVD flag is 0 and other conditions are satisfied, the condition for the BDOF is determined to be satisfied,

When the SMVD flag is 1, it is determined that the condition for the BDOF is not satisfied.
The method of claim 8,

When the condition for the BDOF is satisfied, the prediction sample of the current block is based on a first prediction sample according to the first direction prediction, a second prediction sample according to the second direction prediction, and a BDOF offset in the current block. Is determined,

The BDOF offset is derived based on a first horizontal gradient and a first vertical gradient for the first prediction sample, and a second horizontal gradient and a second vertical gradient for the second prediction sample. .
The method of claim 8,

If the condition for the BDOF is not satisfied, the prediction sample is determined based on a weighted sum of a first prediction sample according to the first direction prediction and a second prediction sample according to the second direction prediction in the current block. An encoding method characterized by.
The method of claim 7,

If the SMVD flag is 1,

The size of the second MVD is the same as the size of the first MVD,

The encoding method, wherein the code of the second MVD is opposite to that of the first MVD.
The method of claim 7,

If the SMVD flag is 1,

The first reference picture for the first direction prediction is determined as a previous reference picture closest to the current picture in display order in the first reference picture list for the first direction prediction,

And the second reference picture for second direction prediction is determined as a next picture closest to the current picture in the display order in the second reference picture list for the second direction prediction.
In the video signal decoding apparatus for inter prediction,

A memory for storing the video signal;

A processor coupled to the memory and processing the video signal,

The processor,

Acquires a symmetric motion vector difference (SMVD) flag of the current block, wherein the SMVD flag is whether a second MVD for second direction prediction is determined based on a first motion vector difference (MVD) for first direction prediction Is related to,

Checking the condition for a bi-directional optical flow (BDOF) based on the SMVD flag,

And generating a prediction sample of the current block based on the condition for the BDOF.
An apparatus for encoding a video signal for inter prediction, comprising:

A memory for storing the video signal;

A processor coupled to the memory and processing the video signal,

The processor,

Generate prediction information for the current block,

It is set to generate a prediction sample of the current block based on the prediction information,

The prediction information includes a symmetric motion vector difference (SMVD) flag related to whether a second MVD for second direction prediction is determined based on a first motion vector difference (MVD) for first direction prediction,

In order to generate the prediction sample, the processor,

Checking the condition for a bi-directional optical flow (BDOF) based on the SMVD flag,

The encoding apparatus, characterized in that the encoding apparatus is set to generate a prediction sample of the current block based on the condition for the BDOF.
In a non-transitory computer-readable medium storing one or more instructions, the one or more instructions executed by one or more processors may cause a video signal processing apparatus to ,

Acquires a symmetric motion vector difference (SMVD) flag of the current block, wherein the SMVD flag is whether a second MVD for second direction prediction is determined based on a first motion vector difference (MVD) for first direction prediction Is related to,

Checking the condition for a bi-directional optical flow (BDOF) based on the SMVD flag,

And controlling a video signal processing apparatus to generate a prediction sample of the current block based on the condition for the BDOF.